Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 1 of 96

IN THE UNITED STATES DISTRICT COURT

FOR THE DISTRICT OF COLUMBIA

HELENA WORLD CHRONICLE, LLC CLASS ACTION COMPLAINT

Plaintiff, JURY TRIAL DEMANDED
v.

GOOGLE LLC and ALPHABET INC.,

Defendants.

COMPLAINT
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 2 of 96

TABLE OF CONTENTS

I. INTRODUCTION ................................................................................................................. 1

II. JURISDICTION AND VENUE ........................................................................................... 5

III. PARTIES................................................................................................................................ 6

A. Plaintiff. ............................................................................................................................... 6
B. Defendants. .......................................................................................................................... 7
IV. AGENTS AND CO-CONSPIRATORS ............................................................................... 7

V. FACTUAL ALLEGATIONS ................................................................................................ 8

A. Concentration and Dominance of Digital Platforms............................................................ 8

B. The Relevant Markets or Lines of Commerce at Issue. ..................................................... 10
1. Google Possesses Monopoly Power And A Dominant Position In The General Search
Market Or Line Of Commerce. ............................................................................................. 10
2. Google Possesses Monopoly Power And A Dominant Position In The Search
Advertising Market. ............................................................................................................... 13
3. Google Possesses Dominant Position In The Line Of Commerce For Digital News And
Reference Searches. ............................................................................................................... 16
C. Anticompetitive Conduct ................................................................................................... 22
1. Acquisitions.................................................................................................................... 22
2. Barriers to Entry. ............................................................................................................ 27
3. Foreclosure Contracts..................................................................................................... 29
4. Tying. ............................................................................................................................. 35
5. Introduction And Implementation of Bard. .................................................................... 55
6. Introduction And Implementation of SGE. .................................................................... 68
7. Changing of Rates in AdSense. ...................................................................................... 75
8. Spoliation. ...................................................................................................................... 76
VI. ANTICOMPETITIVE EFFECTS ..................................................................................... 78

VII. REGULATORY FINDINGS .............................................................................................. 80

VIII. CLASS CERTIFICATION .............................................................................................. 83

IX. CAUSES OF ACTION........................................................................................................ 86

i
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 3 of 96

X. RELIEF ................................................................................................................................ 91

A. DEMAND FOR JUDGMENT........................................................................................... 91

B. JURY DEMAND ............................................................................................................... 92

ii
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 4 of 96

Plaintiff Helena World Chronicle, LLC (“HWC”), on behalf of itself and all other publishers

of original news and reference websites (“Publishers”), 1 brings this Class Action Complaint (the

“Complaint”) against Defendants Google LLC and its parent entity, Alphabet, Inc. (collectively,

“Google”) for violations of: Sections 1, 2, and 3 of the Sherman Act (15 U.S.C. §§ 1-3) and

Section 7 of the Clayton Act (15 U.S.C. § 18). Plaintiff seeks treble damages, injunctive relief, and

damages pursuant to 15 U.S.C. §§ 15 and 26, as a result and consequence of Defendants’ unlawful

conduct. The relevant Class Period extends from November 1, 2019 through the date on which a

Class is certified.

I. INTRODUCTION

1. Google is starving the free press. Every year, it siphons off billions of readers and

billions of dollars from Publishers through an anticompetitive scheme that extracts their content,

publishes it on Google, and diverts readers and ad revenue. This scheme is part of an unlawful

strategy to attract, trap, and maximize users within a “walled garden” that entrenches Google’s

monopoly as the world’s largest search engine. Google has coerced Publishers into a Hobson’s

choice: surrender their content or disappear from search and lose the single largest source of

referral traffic and associated revenue.

2. The key to this scheme is Google’s structure and dominance as a digital platform.

Digital platforms serve as gatekeepers to online markets by leveraging key barriers to entry: scale,

network effects, and switching costs. No platform has greater power than Google. That power now

rests on Google’s monopoly over general search services, where it controls nearly 90% of the

1
The term “Publishers” as used herein refers to publishers of original news and reference websites that produce and
publish original, non-fiction content in digital format, including digital newspapers, magazines, blogs, weather
reports, opinion and editorials, and reference works.

1
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 5 of 96

market and which it monetizes through search advertising. Google’s search monopoly is at the

heart of its dominance as a digital platform. It abuses that dominance to lessen competition in

associated and dependent lines of commerce—including digital news and reference publishing.

3. Google maintains its search monopoly and abuses it dominance through various

tying arrangements that lock in and exploit Publishers as input suppliers; through foreclosure

contracts that block rival search engines, and through more than 260 mergers and acquisitions that

enable it to entrench and enlarge its dominant position.

4. First, Google attracts users through unlawful tying arrangements, where it

misappropriates Publishers’ content and re-publishes it on Google’s platform. The tying product

is Google’s general search service. The tied product is digital news and reference content published

on Google Search through its patented question-and-answer technologies. Google launched this

tying scheme in 2012 through its WebAnswers project. In 2023, it amplified this scheme using

generative artificial intelligence (“GAI”) (in the form of its Bard chatbot, which provides answers

to a user’s questions in natural language) and its Search Generative Experience (“SGE”), which

responds to a user’s search by directing him or her to its own summary of what other websites say.

Google has repeatedly abused its dominance to force Publishers to supply the inputs for this tying

arrangement: their content. It falsely told Publishers that its modifications to Google Search would

increase their referral traffic and revenue, but the truth was the opposite. As one of Google’s AI

engineers recently admitted: “Direct answers reduce search referral traffic” and “[p]ressure”

Publishers to “develop alternative revenue streams.” 2

5. Second, Google traps users through a series of exclusionary agreements with

Apple, Samsung and other mobile device manufacturers and browser developers. These

2
Marc Najork, Generative Information Retrieval, ACM Digital Library (July 24, 2023),
https://dl.acm.org/doi/abs/10.1145/3539618.3591871.

2
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 6 of 96

foreclosure contracts secure Google Search as the default search engine on the vast majority of

mobile devices and desktop computers. In 2002, Google and Apple entered into an anticompetitive

agreement where Google has in recent years paid Apple an estimated $18 billion a year to make

Google Search the default search engine in Safari, Apple’s web browser. As part of this restraint

on trade, Google currently pays Apple 36% of its search revenue over the Safari browser. Google

has made similar agreements with Samsung, the manufacturer of mobile devices running Google’s

Android operating system. Coupled with Google’s bundling Google Search with its own products,

including the Google Chrome browser, these agreements protect and maintain Google’s nearly

90% market share in general search services.

6. Third, Google maximizes this anticompetitive scheme through more than 260

mergers and acquisitions, enabling it to abuse and extend its dominance in, among others, four key

areas: digital display advertising, digital content, mobile operating systems, and AI. For example,

Google’s acquisition of DoubleClick in 2008 enabled it to dominate digital display advertising on

Publishers’ websites. Google’s acquisition of Android in 2005 enabled it to foreclose competition

in search and cement its position as the default search engine on mobile devices. Google’s

acquisition of YouTube in 2006 gave it the largest trove of video content on the internet and

allowed it to engage in self-preferencing of YouTube over other Publishers’ content. Google’s

acquisition of DeepMind Technologies in 2014—and its 2023 merger of DeepMind with Google’s

own AI research lab Google Brain—is enabling it to protect its search monopoly from potentially

disruptive GAI technology.

7. The anticompetitive effects of Google’s scheme cause profound harm in its

monopoly markets to digital news and reference Publishers—and ultimately to the marketplace of

ideas. Google has created a zero-click world in which users remain in its ecosystem and are

3
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 7 of 96

siphoned away from Publishers. Already, 69% of all Google searches result in zero click-through

traffic. Since roughly 80% of all Google searches are for informational content (i.e., news and

reference content), billions of consumers are now getting their news from Google. With an

estimated audience of 69.6 billion information-consumers per month, Google.com is far and away

the largest publisher of online news and reference content.

8. Google does not produce this content; Publishers do. But they are forced to compete

on an unlevel playing field against their own products. The result is a staggering harm to

competition, to consumers, to labor and to a democratic free press.

9. Publishers’ costs rose and their revenue fell from roughly $50 billion in 2005 to

$20 billion in 2022. Since 2005, America has lost more than a fourth of its newspapers (2,500) and

is expected to lose a third by 2025. 3 Only one publisher of encyclopedias—World Book—remains

in print. Seventy million Americans now live in news deserts, with little to no local news coverage. 4

Newspaper employment has declined by 70% from 2005-22. Fewer reporters mean less original

reporting. Fewer editors and fact checkers mean fewer safeguards against misinformation.

Consumers face a loss of quality and choice in the marketplace of ideas. And America faces the

potential collapse of a vital pillar of democracy.

10. Plaintiff HWC owns and publishes two weekly local papers in Arkansas. It cannot

afford to sacrifice the referral traffic it receives from Google. But it also cannot afford to sustain

competitive injury under Google’s search monopoly and abuse of dominance. The antitrust laws

were designed to prevent a monopolist, or a structure achieved in part, through merger and

acquisition, from harming competition by abusing its monopoly or dominance. Plaintiff invokes

3
Penny Abernathy, The State of Local News: The 2022 Report, Northwestern Local News Initiative (June 29, 2022),
https://localnewsinitiative.northwestern.edu/research/state-of-local-news/report/.
4
Id.

4
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 8 of 96

the Sherman Act and Clayton Act to seek classwide monetary and injunctive relief to restore and

ensure competition for digital news and reference publishing, protect Publishers, and set up

guardrails to preserve a free marketplace of ideas in the new era of artificial intelligence.

JURISDICTION AND VENUE

11. This Court has subject matter jurisdiction pursuant to Section 4 of the Sherman Act,

15 U.S.C. § 4, and 28 U.S.C. §§ 1331 and 1337(a).

12. This Court has personal jurisdiction over each Defendant because Defendants do

extensive business within this District — including by providing the monopolized services to class

members and consumers within this this district—and this action arises out of Defendants’ contacts

within this District.

13. Venue is proper in this Judicial District pursuant to Sections 4 and 12 of the

Clayton Act (15 U.S.C. §§ 15 and 22), and 28 U.S.C. § 1391, because: (a) each Defendant transacts

business and is found within this District; (b) a substantial part of the events giving rise to the

alleged claims occurred in this District; and (c) a substantial portion of the affected interstate trade

and commerce was carried out in this District. Each Defendant has transacted business, maintained

substantial contacts, and/or committed overt acts in furtherance of the illegal scheme and

conspiracy throughout the United States, including in this District. Defendants’ conduct has had

the intended and foreseeable effect of causing injury to persons residing in, located in, or doing

business throughout the United States, including in this District.

14. Defendants’ conduct affects interstate trade and commerce. Defendants’ conduct

has a direct, substantial, and reasonably foreseeable effect on commerce within the United States.

II. JURISDICTION AND VENUE

15. This Court has subject matter jurisdiction pursuant to Section 4 of the Sherman Act,

15 U.S.C. § 4, and 28 U.S.C. §§ 1331 and 1337(a).

5
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 9 of 96

16. This Court has personal jurisdiction over each Defendant because Defendants do

extensive business within this District — including by providing the monopolized services to class

members and consumers within this this district—and this action arises out of Defendants’ contacts

within this District.

17. Venue is proper in this Judicial District pursuant to Sections 4 and 12 of the Clayton

Act (15 U.S.C. §§ 15 and 22), and 28 U.S.C. § 1391, because: (a) each Defendant transacts business

and is found within this District; (b) a substantial part of the events giving rise to the alleged claims

occurred in this District; and (c) a substantial portion of the affected interstate trade and commerce

was carried out in this District. Each Defendant has transacted business, maintained substantial

contacts, and/or committed overt acts in furtherance of the illegal scheme and conspiracy

throughout the United States, including in this District. Defendants’ conduct has had the intended

and foreseeable effect of causing injury to persons residing in, located in, or doing business

throughout the United States, including in this District.

18. Defendants’ conduct affects interstate trade and commerce. Defendants’ conduct has

a direct, substantial, and reasonably foreseeable effect on commerce within the United States.

III. PARTIES
A. Plaintiff.

19. Plaintiff HWC is a Publisher. It owns two newspapers: Helena World and Monroe

County Argus, the latter of which was acquired in 2022. Both newspapers are available in print

and digital media and both deliver news and information to subscribers and online readers. Helena

World is printed out of Helena, Arkansas. Over the past 365 days, Helena World (online) has

averaged 1530 visitors per website per month with 402 referred from Google Search per month. The

Monroe County Argus is printed in Brinkley, Arkansas. It has been in circulation since 1875 and

6
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 10 of 96

now is also available online. The online version of Monroe County Argus likewise receives visits

referred from Google Search. HWC’s original works have been scraped and reproduced for the

purpose of training Google’s GAI products including Bard and SGE, and Bard has since plagiarized

an article from Helena World. On a daily basis, HWC’s original works continue to be scraped and

reproduced as inputs by Google’s GAI search product and are plagiarized on a regular basis as

Artificial Intelligence (“AI”)-generated news summaries and extracts.

B. Defendants.

20. Defendant Google LLC is a limited liability company organized and existing under the

laws of the State of Delaware, with its principal place of business in Mountain View, California.

Google LLC is an online advertising company providing internet-related products, including various

online advertising technologies, directly and through subsidiaries and business units that it owns and

controls.

21. Defendant Alphabet Inc. is a publicly traded company incorporated and existing under

the laws of the State of Delaware and headquartered in Mountain View, California. Alphabet Inc.

was created as a holding company for Google in late 2015, and Alphabet controls Google’s day-to-

day operations. Virtually all of Alphabet Inc.’s revenue comes from Google LLC. Since December

2019, Alphabet and Google have had the same Chief Executive Officer (Sundar Pichai (“Pichai”)).

As a result of Alphabet Inc.’s operational control, Google LLC is Alphabet Inc’s alter ego.

22. Google LLC and Alphabet Inc. are referred to herein collectively as “Google.”

IV. AGENTS AND CO-CONSPIRATORS

23. The unlawful acts of Defendants set forth in this class action complaint were

authorized, ordered, or performed by the Defendants’ respective officers, agents, employees,

representatives, or shareholders while actively engaged in the management, direction, or control of

the Defendants’ businesses or affairs. The Defendants’ agents operated under the explicit and apparent

7
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 11 of 96

authority of their principals. Each Defendant, and its subsidiaries, affiliates, and agents operated as a

single unified entity.

24. Various persons and/or firms not named as Defendants may have participated as co-

conspirators in the violations alleged herein and may have performed acts and made statements in

furtherance thereof. Each acted as the principal, agent, or joint venture of, or for other Defendants

with respect to the acts, violations, and common course of conduct alleged herein.

V. FACTUAL ALLEGATIONS

A. Concentration and Dominance of Digital Platforms.

25. Today’s digital economy is highly concentrated in the hands of a few firms that

dominate online markets and lines of commerce. These “digital platforms” operate as essential

mediation services, control key channels of distribution, and act as gatekeepers between consumers,

Publishers, and advertisers.

26. On October 30, 2023, President Joe Biden issued an Executive Order on the

development and use of AI in the United States. 5 One of the points that he emphasized was the need

to “promote competition in AI and related technologies, as well as in other markets. Such actions

include addressing risks arising from concentrated control of key inputs, taking steps to stop

unlawful collusion and prevent dominant firms from disadvantaging competitors….” 6 (Emphases

added). As the House Antitrust Subcommittee’s 2020 Report concluded, “dominant platforms are

able to exploit their gatekeeper power to dictate terms and extract concessions that no one would

5
The White House, Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial
Intelligence, 88 Fed. Reg. 75191 (Oct. 30, 2023), https://www.whitehouse.gov/briefing-room/presidential-
actions/2023/10/30/executive-order-on-the-safe-secure-and-trustworthy-development-and-use-of-artificial-
intelligence/.
6
Id.

8
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 12 of 96

reasonably consent to in a competitive market.” 7

27. The Federal Trade Commission (“FTC”) recently expressed similar concerns, saying

that the use of GAI raises significant competition issues by, for example, “further entrenching the

market power” of dominant firms. 8 There is no precedent in the United States economy for the power

of today’s digital platforms. And there is no digital platform with a greater sphere of influence than

Google. This lawsuit is intended to address such a failure of competition that exists now with respect

to the digital news and reference publishing aspect of Google’s platform.

28. In 2020, the Antitrust Division of the United States Department of Justice (“DOJ”)

and various state attorneys general sued Google for antitrust violations. See United States v.

Google LLC., No. 1:20-cv-03010-APM (D.D.C.) (“DC DOJ Case”). Many of the Google internal

documents referenced throughout this Complaint have been used publicly at that trial, which ended

on November 16, 2023. 9

29. In analyzing Google’s unlawful maintenance of monopoly power and abuse of its

dominant position in the market (or line of commerce under Section 7 of the Clayton Act) for internet

general searches and the market (or line of commerce) for advertising searches, Google’s conduct

must be viewed holistically over the last two decades.

30. The facts set forth below summarize Google’s monopoly power and dominance within

7
Subcommittee on Antitrust, Commercial and Administrative Law of the Committee on the Judiciary, Investigation
of Competition In Digital Markets, Majority Staff Report. at 11 (2020) (“House Subcommittee Report”).

8
FTC, In Comment Submitted to U.S. Copyright Office, FTC Raises AI-related Competition and Consumer
Protection Issues (Nov. 7, 2023), https://www.ftc.gov/news-events/news/press-
releases/2023/11/InCommentSubmittedtoUSCopyrightOfficeFTCRaisesAIrelatedCompetitionandConsumerProtecti
onIssuesStressingThatItWillUseItsAuthoritytoProtectCompetitionandConsumersinAIMarkets.
9
Although the DC DOJ Case did not include any extensive analysis of Google’s more recent conduct regarding
Bard and SGE, and does not focus on Publishers, the trial was replete with expert testimony on some of the same
dominant market power that is at issue here and underscores the fact that abuse of that market power can have
devastating anticompetitive impact in the digital world.

9
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 13 of 96

the markets or lines of commerce at issue. That power and dominance was created in part through

key acquisitions of other companies; contracts with device manufacturers like Apple and Samsung

that excluded or foreclosed competing search engines from gaining a foothold; a 2016 contract

between Google and Apple that prevented the latter from developing any competing search engine

of its own; the years-long effort by Google to contain users seeking news and reference content in

its “walled garden” ecosystem; the misappropriation of Publishers’ web content; the rushed rollout

of Bard in 2023 in order to stifle competition from Microsoft; the introduction of SGE; and Google’s

decision this year to modify its AdSense terms in a way that disadvantages Publishers. 10 Google’s

abuse of its monopoly power or dominance has harmed Publishers in both the United States and

abroad. Each of these topics is discussed separately in the sections that follow.

B. The Relevant Markets or Lines of Commerce at Issue.

1. Google Possesses Monopoly Power And A Dominant Position In The

General Search Market Or Line Of Commerce.

31. Google was launched in 1998 as an internet search engine, serving search results

linking to third-party websites in response to users’ queries. 11 Google’s key feature was the algorithm

PageRank, which ranked the relevance of a webpage to a given search query by examining how

many other webpages linked to that page. By 2000, Google had become the largest search engine in

the world. 12 Its “ten blue links” became the gateway to the web for billions of users. The brand name

“Google” became a verb synonymous with internet search itself. When Google was in its infancy,

observers questioned how Google could monetize what appeared to be a free internet search

10
Google AdSense is a program designed for website publishers who want to display specific text, video or image
advertising on pages of their website and make money when visitors to the site see or click on ads.

11
Google Inc., Registration Statement (Form S-1), at 1 (Apr. 29, 2004),
https://www.sec.gov/Archives/edgar/data/1288776/000119312504073639/ds1.htm.
12
House Subcommittee Report at 174.

10
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 14 of 96

service. The answer was advertising. Google adopted a triple-product business model:

• Google Search gives users access to online content;

• Google collects user attention and data; and

• Google sells audience attention and data to advertiser 13

To become the behemoth that it is today, Google needed to maximize three key things: content,

users, and advertising. It did this by leveraging key barriers to entry to acquire and maintain a

monopoly in the market for general search services and entrench itself as a dominant digital

platform.

32. The general search services market (or line of commerce for Clayton Act purposes)

consists of “general search engines, which are ‘one-stop shops’ consumers can use to search the

internet for answers to a wide range of queries.” 14 General search engines can answer all types of

queries and return a wide range of results.

33. By contrast, what are referred to as “vertical search engines” are limited to specific

topics. For example, a search on Google for “sore throat treatment” returns a range of results, from

medical products to the latest scientific information and tips on treatment. However, a search on

Amazon—a vertical search service provider—only returns results for medical products. Consumers

would not find a vertical search a suitable substitute for general search.

34. The relevant geographic scope of the general search market is the United States.

Google provides users in the United States a distinct website that differs from those provided by

Google in other countries.

13
See Carlos Diaz Ruiz, Disinformation on digital media platforms: A market-shaping approach, New Media &
Society 9-10 (Oct. 30, 2023), https://journals.sagepub.com/doi/10.1177/14614448231207644.

14
Mem. Op. at 2, United States, et al. v. Google LLC, No. 1:20cv03010 (D.D.C. Oct. 6, 2023), ECF No. 728 at 6
(“SJ Op.”).

11
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 15 of 96

35. By any reasonable assessment, Google possesses monopoly power in the United States

market (or line of commerce pursuant to Section 7 of the Clayton Act) for general searches. According

to the website Statcounter: 15

36. As shown above, Google’s share of this market or line of commerce has been consistently

just below 90%. By contrast, Microsoft Corporation’s (“Microsoft”) market share, through Bing, which

incorporated the ChatGPT chatbot discussed below, has been consistently minimal in recent years. In

fact, according to one source, Microsoft’s market share for Bing declined in 2023, relative to 2022.16 In

denying summary judgment for the defense, the court in the DC DOJ Case emphasized Google’s

dominance of the general search services market: “[t]here are other search engines, of course: Microsoft’s

15
Statcounter: Global Stats, Search Engine Market Share United States of America, Nov. 2022 – Nov. 2023,
https://gs.statcounter.com/search-engine-market-share/all/united-states-of-america (last accessed Oct. 29, 2023).

16
Danny Goodwin, The new Bing has failed to take any market share from Google after six months, Search Engine
Land (Aug. 17, 2023), https://searchengineland.com/new-bing-google-market-share-six-months-430840.

12
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 16 of 96

Bing, Yahoo!, and DuckDuckGo, to name a few. But their market penetration pales in comparison to

Google’s. In 2020, Google’s share of the U.S. general search services market was nearly 90%, and even

higher on mobile devices. The market share of Google's closest competitor, Bing, was roughly 6%.”17

37. Michael Whinston (“Whinston”), a Professor at the Massachusetts Institute of

Technology who testified as an expert for the government in the DC DOJ Case, stated that if one

considered searches conducted through mobile phones, Google’s market share was just under

95%. 18He also said that there were high barriers to entry that allowed Google to sustain its market

power over a long period of time. These included fixed and sunk costs of operation, Google’s brand

recognition, the scale of its operations, and its control of access points through contracts with device

makers, which will be described below. 19

2. Google Possesses Monopoly Power And A Dominant Position In The

Search Advertising Market.

38. Google also has monopoly power and a dominant position in the search advertising

market. 20 Whinston estimated that as of 2021, Google has approximately 74% share of that market

as of 2020. 21 Reports in 2023 variously indicate that Google has a 79.8% or 71% share of the

17
SJ Op. at 2. The Court further observed that “because of its large market share in general search services, Google
also holds a superior market position in various search-related advertising market.” Id.

18
Trial Tr., United States, et al. v. Google LLC, No. 1:20cv03010 (D.D.C. Oct. 6, 2023) (Day 18) at 4762:25-
4763:2.

19
Id. at 4764:12-4765:3, 4766:2-9, 4766:13-15; 4767:7-10, 15:16.

20
Whinston also identified what was essentially a submarket of general search consisting of general text advertising
searches, in which Google had roughly the same market share as it does in the general search market. Id. at 4777:10-
13. That will not be discussed in this Complaint because it part of the broader market or line of commerce for
advertising searches generally.

21
Id. at 4779:15.

13
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 17 of 96

advertising search market in the United States. 22 More than 80% of businesses worldwide rely on

Google ad campaigns. 23 This power and position in the search advertising market is driven in part

by Google’s control over the general search market.

39. Google’s dominance in the search advertising market bears upon Plaintiff’s claims

involving and Google’s recent change in how it compensates Publishers for AdSense traffic, which

is described in detail below. Its dominance is also a factor in its ability to engage in self-preferencing

tying conduct.

40. In 2000, Google branched out from search to digital advertising, launching AdWords,

its ad buying platform for such advertising. This service enabled businesses to purchase keyword

advertising that would appear on Google’s Search Engine Results Page (“SERP”). This launch

marked the first step in Google’s transformation of its SERP—the world’s portal to the web—into a

medium for publishing content and ads, in addition to hyperlinks.

41. Since 2000, Google has increased its power as a digital platform and expanded beyond

search into numerous lines of commerce. It has become “the largest provider of digital advertising,

a leading web browser, a dominant mobile operating system, and a major provider of digital

mapping, email, cloud computing, and voice assistant services, alongside dozens of other

offerings.” 24

42. It was estimated that in 2020 alone, Google made $147 billion in revenue from online

ads and “[a]bout 16% of its revenue came from the company’s display or network business, in which

22
See Ahivanjali Pawar, Most Vital Google Ads Statistics Traders Need to Grasp in 2023 and Beyond, Enterprise
Apps Today (Oct. 5, 2023), https://www.enterpriseappstoday.com/stats/google-ads-statistics.htmll; Statista, Search
Advertising-United States. https://www.statista.com/outlook/amo/advertising/search-advertising/united-states (last
accessed Oct. 30, 2023).

23
Id.

24
House Subcommittee Report at 174.

14
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 18 of 96

other media companies use Google technology to sell ads on their website and apps.” 25 In 2022, as

described in further detail below, the European Publishers Council (which includes Press Publishers

like Axel Springer, News UK, Conde Nast, Bonnier News, and Editorial Prensa Iberica) sued Google

over its internet ad search advertising practices. 26

43. Internet search advertisements are distinct from general searches. In the words of Hal

Varian (“Varian”), the Chief Economist at Google, search advertisements satisfy demand rather than

create it. 27 For example, Booking.com made the following distinction between the two as shown in

the graphic below. 28

25
Foo Yung Chee, Google’s advertising tech targeted in European publishers’ complaint, Reuters (Feb. 11, 2022),
https://www.reuters.com/technology/googles-advertising-tech-targeted-european-publishers-complaint-2022-02-11/.

26
Id.

27
Ex. No. UPXD103, United States, et al. v. Google LLC, No. 1:20cv03010 (D.D.C.),
https://www.justice.gov/d9/2023-10/417081.pdf at 22 (last accessed Oct. 30, 2023).

28
Id. at 23.

15
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 19 of 96

44. Apple recognized a similar point in January of 2017: 29

Advertisers recognize that Google is a necessary source of services for internet search

advertising and that any meaningful alternative is limited. 30

3. Google Possesses Dominant Position In The Line Of Commerce For

Digital News And Reference Searches.

45. Google exercises its dominance in the line of commerce for digital news and reference

publishing.

46. News search is the single largest source of external referral traffic to news websites,

exceeding referral traffic from social media and direct traffic to publishers’ websites. 31 Users of news

search services are a distinct audience, with distinct demands from those consuming news through

29
Id.

30
Id. at 30, 31.

31
Aisha Majid, Search vs. social: How referral traffic to news sits has changed in five years, Press Gazette (April
13, 2023) (“Press Gazette Article”),
https://pressgazette.co.uk/media-audience-and-business-data/media_metrics/news-referral-traffic-breakdown/

16
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 20 of 96

other distribution channels. According to one study, 69% of news-search users who click through to

news websites only consume one category of news: “[b]ecause Search audiences are drawn to stories

with factual, timely information, they are less likely to move on to other categories after satisfying

their original inquiry.” 32

47. Digital news and reference publishing is a significant line of commerce in the United

States. Before the internet became ubiquitous, Americans obtained original, fact-based informational

content from newspapers, magazines, and reference works (encyclopedias, dictionaries, almanacs,

etc.). Since the emergence of the web, the news, media, and reference publishing industry has shifted

to online distribution and consumption, as part of the digital content media ecosystem. The

percentage of Americans getting at least some of their news online rose from 2% in 1995 to 93% in

2018. 33 Today, only one publisher of encyclopedias—World Book—remains in print. Digitization

has led to the development of new formats for publishing informational content online, including

digital magazines and newspapers, encyclopedias, wiki, blogs, news aggregators, news search

platforms, and digests. In 2001, in recognition of this distinct emerging market, a trade association

called the Online Publishers Association (now called Digital Content Next) was formed, with

members including print and digital newspaper publishers (e.g., The Atlanta Journal-Constitution,

New York Times) and specialized vertical-content publishers (e.g., AARP, AccuWeather, Consumer

Reports, WebMD). 34 Digital news is a key line of commerce within this industry. In the United

32
Jack Neary, How widely do visitors read across news and media sites? Our data on breadth of visit, Chartbeat
https://blog.chartbeat.com/2023/08/01/how-widely-do-visitors-read-across-news-and-media-sites/ (last accessed
Oct. 30, 2023).

33
News Media Alliance, Google Benefit from News Content: Economic Study at 3 (June 2019),
https://www.newsmediaalliance.org/wp-content/uploads/2019/06/Google-Benefit-from-News-Content.pdf.

34
Kendall Moe, Choose your words wisely: The role of language in media trust, Digital Content News (Nov. 16,
2023), https://digitalcontentnext.org/blog/2023/11/16/choose-your-words-wisely-the-role-of-language-in-media-
trust/.

17
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 21 of 96

States, digital news publishing revenue is expected to exceed $27 billion by 2027. 35

48. Vertical news and reference search services encompass the breadth of non-

commercial “informational” searches which make up 80% of Google’s total searches. 36 Google

Search’s website, google.com, is the most widely used digital news and media website in the world,

by a landslide: Google receives an estimated 69.8 billion non-commercial, informational search

queries per month. In contrast, Wikipedia (Google’s largest competitor in digital news and media

publishing) received only 4.5 billion visits in October of 2023. 37 In comparison, the largest

traditional news Publisher webpages receive monthly visits in the tens or hundreds of millions, not

billions. 38 In addition to Google Search’s news surfaces, Google News (a news aggregator website

and app) had 398.1 million visits in October 2023—the 16th most popular news site in the world. 39

49. Not only does Google own the most widely used search engine in this country, but it

functions as a news provider itself. Google plays several roles in the digital informational media

ecosystem: (1) connecting consumers to Publishers’ websites, (2) selling targeted search advertising,

(3) facilitating display advertising on Publishers’ websites, and (4) publishing Publishers’ news and

reference content on Google’s webpages and apps. This means that Google maintains both vertical

and horizontal relationships with Publishers. In their vertical relationship, Google connects

Publishers to users as an intermediary. Specifically, Google “sells” search-traffic referral services to

Statista, U.S. digital publishing industry—statistics and facts (Aug. 30, 2023),
35

https://www.statista.com/topics/1453/digital-publishing/#topicOverview.
36
Id.

37
Similarweb.com, Website Performance: Wikipedia.com (Oct. 2023).

38
The Wall Street Journal has roughly 83.5 million unique monthly visitors. Improvado, Wall Street Journal Media
Kit, https://improvado.io/resources/wall-street-journal-media-kit (last accessed Oct. 30, 2023). The New York Times
has roughly 130 million monthly readers. Maria Pengue, 25 Insightful New York Times Readership Statistics – The
2023 Edition, WorkUp (Mar. 14, 2021), https://letter.ly/new-york-times-readership-statistics/.

39
Similarweb.com, News and Media Category Leaders (Oct. 2023).

18
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 22 of 96

Publishers, which they purchase by giving Google access to their informational content.

50. Google thus competes with Publishers horizontally, as a publisher of informational

content in its own right. In a competitive market, they must either produce or purchase the inputs:

informational content. But Google leverages its search engine monopoly to do neither. It

misappropriates third-party Publisher content, without paying licensing fees. This raises Publishers’

costs by forcing them to bear the labor and overhead costs of producing informational content plus

the extortionate price of Google’s search-referral traffic service.

51. Put differently, Google’s conduct artificially suppresses the revenue flowing to

Publishers, who are no longer compensated with a fair flow of users’ seeking the content they

purchase or produce. Those users stay in Google’s “walled garden” ecosystem, where Google feeds

users the misappropriated publisher content directly. Google thus maintains its monopoly in search

by maximizing user attention on Google’s platform: users who do not navigate away from Google’s

SERP spend more time on Google’s platform, provide more data to Google, and perform more

searches, thereby exposing them to targeted search advertising. At the same time, Google uses the

misappropriated content to increase its share and dominance in the news and reference content

publishing line of commerce.

52. News search services are unique because they provide search results cross-ranked by

commercial segment (i.e., news), relevance, and timeliness. From a consumer perspective, the line

of commerce for vertical news search is distinct from the market for general search or other verticals.

For example, if a consumer searches for “Barack Obama” in a general search engine, the consumer

will expect a variety of results to be returned, including commercial results, as well as older and

evergreen results: e.g., U.S. government websites, biographical information about Barack Obama’s

life, links to his personal webpage and social media accounts, images and videos of the former

19
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 23 of 96

President, encyclopedia entries, etc. The results might originate from a combination of social media,

governmental agency websites, academic articles, Wikipedia, and beyond.

53. On the other hand, a user searching for “news about Barack Obama” would expect to

receive the latest information on the topic and to only see products from news publishers. This

tailored service is precisely what Google Search provides when a user clicks on the “News” vertical

and switches from general search to vertical news search. Indeed, in contrast to general search results,

vertical news search results are ranked chronologically to meet a user’s distinct demand for timely

information in this search market.

20
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 24 of 96

54. Google also displays news digests and reference content across what it calls “News

surfaces” and “Knowledge Panels” across its various webpages and apps:

• Google Search Top Stories Box (News Carrousel)

• Google Search Perspectives Box

• Google Search Vertical News Tab

• Google Search “People also ask” Box

• Google Search “About” Box (Knowledge Graph)

• Google Search “Knowledge Panels” (Knowledge Graph)

• Google Search Generative Experience (chatbot integrated into Google Search

reproducing news articles and answering informational queries)

• Google News (news aggregator website and app)

• Google Discover (recommended social feed)

• Google Bard (chatbot summarizing news articles and answering informational

queries)

• YouTube News Channel

21
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 25 of 96

55. Google’s dominance over vertical news search services is underscored by its 95%

share of all referral traffic to news websites from its search engines originates from Google, while

only 5% originates from a handful of competitors: Bing News, Yahoo News!, and DuckDuckGo

News. 40

C. Anticompetitive Conduct

1. Acquisitions.

56. Google began its march to dominance in the aforementioned markets and lines of

commerce by acquiring key competitors in mobile devices, digital media, AI, and digital

advertising.

57. Google achieved its structure as a dominant digital platform, through a series of

strategic mergers and acquisitions designed to attract, trap, and monetize users of its search engine

services.

58. To date, Google has acquired 260 different entities, across various lines of commerce,

each in furtherance of its walled garden scheme.

59. Several acquisitions highlight Google’s commitment to fortifying its walled garden in

search, through the acquisitions of entities in related and dependent lines of commerce. Four of these

acquisitions—of the mobile start-up Android, the video platform YouTube, the AI firm DeepMind,

and of the digital ad vendor DoubleClick—both reinforce Google’s monopolies and enable,

entrench and extend its overall digital platform dominance:

60. YouTube. Google acquired YouTube, the digital video platform, in 2006, in

furtherance of its “attract” strategy, to draw users to its platform with video content. Not only did the

acquisition of YouTube serve to attract users to the Google product family in general, but also

40
Press Gazette Article.

22
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 26 of 96

Google specifically leveraged the YouTube acquisition to fortify and contain users in its walled

garden ecosystem in search. News search queries on Google’s SERP frequently display a “Videos”

panel, which links directly to YouTube, regardless of whether a different third-party publisher

originally served the YouTube video from a news article or webpage outside of YouTube. Thus,

even if users click through for more information, they are not even taken to Publishers’ websites;

instead, they are diverted by Google Search into one of Google’s platforms for publishing digital

news: YouTube.

23
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 27 of 96

61. DeepMind. Google’s acquisition of AI firm DeepMind Technologies, in 2014

furthers Google’s ability to both attract and trap users in its walled garden search engine. Google’s

acquisition of DeepMind and other AI firms, because this technology enabled Google to further

refine its search engine algorithm, by enabling Google to extract more value out of the data Google

24
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 28 of 96

collected from the billions of users it had attracted to and trapped on its platform. 41 Moreover, as

discussed above, AI, when integrated into Google’s SERP as a means of publishing news content,

will further attract users to Google’s walled garden search engine by enabling Google to publish to

users exponentially more of the news content Google misappropriates from news publishers, without

users ever needing to leave Google’s SERP. In other words, Google’s AI technologies—developed

in part by its mergers and acquisitions—will strengthen Google’s ability to tie its general search

services to its digital news publishing services.

62. Android. In 2005, Google’s acquired Android, today the world’s dominant mobile

operating system running on 75% of the world’s mobile devices. As detailed in the section above,

Google leveraged its acquisition of Android to extend its search monopoly, and increasingly “trap”

users within its walled garden ecosystem. A recent House Subcommittee on Antitrust Report

captured Google’s intention to use its purchase of Android to strengthen and entrench its search

monopoly, noting that: “Google used its search engine dominance and control over the Android

operating system to grow its share of the web browser market and favor its other lines of

business.” 42

63. DoubleClick. Google acquired digital ad vendor DoubleClick in 2007, in furtherance

of its scheme to “monetize” its search monopoly. While Google views most of its informational news

searches as “advertising” to attract users to its walled garden, Google earns the majority of hundreds

of billions of dollars of yearly revenue through searches which it monetizes through search

advertising. Google’s acquisition of DoubleClick and related advertising technologies enable Google

to extract revenue from the billions of users it has attracted to and trapped on its search engine.

41
See Pl’s Pre-Trial Br., at 10, United States, et al. v. Google LLC, No. 1:20cv03010. (“Large-scale machine
learning,” Google posits, reveals that “the more we learn form our users, the better we can serve them.”)

42
House Subcommittee Report at 225

25
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 29 of 96

Despite promising regulators that it would not it would not combine the data collected on internet

users via DoubleClick with the data collected throughout Google’s ecosystem, not even a decade

later, Google broke that promise “effectively combining information from a user’s personal identity

with . . . their search history” and other data collected from various Google products.

64. Other Acquisitions. In addition to these four acquisitions, the DOJ has identified

three other key acquisitions that cemented Google’s dominance. These were: (a) the acquisition

for $750 million in 2009 of AdMob, a technology system that allows publishers of mobile apps to

sell ads; (b) the acquisition of InviteMedia, which had a demand site platform, for $81 million in

2010.; and (c) the acquisition in 2011 of AdMeld for $400 million, which had a technology that

provided “yield management” functionality to Publishers. They are described in paragraphs 71-88

of the complaint in a separate case filed by the DOJ against Google in 2023 in the Eastern District

of Virginia (“DOJ Va Case”). 43 As stated in ¶ 88 of the DOJ Va Case complaint, “[t]he DoubleClick,

InviteMedia, and AdMeld acquisitions helped Google achieve dominant positions at each level of

the open web ad tech stack and set the stage for Google to control and manipulate the process by

which publishers sell and advertisers buy open web display inventory.” 44

65. At the time of approval of these acquisitions, it was neither known nor foreseen that

the newly created structure would in fact be used to substantially lessen competition in lines of

commerce related to Google’s general search services monopoly, including in advertising, AI, and

digital new and content publishing. Over time, however, this is exactly how Google entrenched and

43
Compl., United Sates, et al. v. Google LLC , No. 1:23-cv-00108 (E.D. Va. Jan. 24, 2023), ECF No. 1,
https://www.justice.gov/d9/press-releases/attachments/2023/01/24/us_v_google_complaint_0.pdf. (“ED Va
Compl.”)

44
Id.at ¶ 88; See Karina Montoya, How Three Mergers Buttressed Google’s Ad Tech Monopoly, per DOJ, Tech
Policy Press (Mar. 9, 2023), https://techpolicy.press/how-three-mergers-buttressed-googles-ad-tech-monopoly-per-
doj/.

26
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 30 of 96

extended its dominance.

66. Indeed, regulators have admitted as much. For example, as to DoubleClick, former

FTC commissioner William Kovacic explained to the New York Times: “If I knew in 2007 what I

know now, I would have voted to challenge the DoubleClick acquisition.” 45 As United States

regulators have said, “[t]hrough design choices and default settings, Google can use its dominance

in any one market to favor its other lines of business.” 46 And it has done so.

67. Taken together, Google’s series of strategic mergers and acquisitions fortified its

dominance to facilitate and implement anticompetitive conduct across its platform.

2. Barriers to Entry.

68. Google’s dominance as a digital platform is enhanced and protected by several

barriers to entry created in part through these acquisitions. The first barrier to entry is network effects.

In a market with strong network effects, the more people who use a product or service, the more that

product or service becomes useful and valuable. For example, Google’s value as a platform that

facilitates advertising increases as the number of its users grows, because advertisers can reach more

consumers and acquire more user data, enabling more targeted advertising. Such network effects, in

conjunction with other barriers, ensure that Google has enduring market power.

69. A second barrier to entry consists of switching costs. Digital platforms such as Google

can maintain market power because it is difficult, inconvenient, and costly to switch away from

incumbents’ technology and services.

70. A third barrier to entry is economies of scale. In a market characterized by economies

45
Steve Lohr, This Deal Helped Turn Google Into an Ad Powerhouse. Is That a Problem?, New York Times (Sept.
21, 2020), https://www.nytimes.com/2020/09/21/technology/google-doubleclick-antitrust-ads.html.
46
House Subcommittee Report at 226.

27
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 31 of 96

of scale, costs decrease as sales increase. Google benefits from economies of scale in its core service

of providing information. Google’s products—from Google Search to Google Maps and Bard—

have high up-front fixed costs but may scale up with relatively low increases in cost.

71. A fourth barrier to entry is data accumulation. Digital platforms such as Google can

use their massive trove of user data as a barrier to entry against competitors. This is a form of network

effect: platforms with more data can better target advertising and make their services more attractive

to users, drawing more attention and user engagement, which in turn enables more data harvesting.

Digital platforms can leverage data accumulated through their gatekeeping functions to gain

advantages over competitors in other lines of commerce, furthering user lock-in to a range of

products and services.

72. A fifth barrier to entry consists of foreclosure agreements with device makers, as

discussed separately below.

73. As one Google investor, Roger McNamee, was quoted as saying in the House

Subcommittee Report:

Essentially, the interplay of Google’s dominant position in … infrastructure elements

[such as] ad tech infrastructure, Chrome browser, [and Nest] … collectively provide
leverage over other market participants, which include not just startups, but also
advertisers, and other would-be competitors. And the key thing is, it’s not just about
Google’s infrastructure. When you add in Gmail, Search, Maps, apps, and all the other
things that Google does so well … [t]hey provide further levels of user lock-in—further
protective modes that really limit the opportunity of competitors and even, frankly,
suppliers and advertisers, to do the things that they should be able to do in a freely
competitive economy. 47

74. Consequently, today, Google is one of the world’s largest corporations: “Nine of

Google’s products—Android, Chrome, Gmail, Google Search, Google Drive, Google Maps, Google

47
House Subcommittee Report at 43-44.

28
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 32 of 96

Photos, Google Play Store, and YouTube—have more than a billion users each.” 48 Google reported

$279.8 billion in revenue in 2022. Although only 20% of queries on Google Search return search

ads, $224.47 billion is derived from advertising revenue. 49

3. Foreclosure Contracts.

75. A critical component of Google’s anticompetitive conduct is its use of foreclosure

agreements to trap users on its platform, thereby capturing their valuable attention and data, and

starving potential competitor-providers of general search services of the opportunity to reach those

users. These foreclosure agreements not only foreclose competition in the search engine market, but

also have anticompetitive reverberations throughout lines of commerce where are related to and

dependent on online search—including digital news publishing.

76. Three key sets of agreements form the basis of this foreclosure: (1) an agreement

with Apple to make Google the default search engine on Apple’s mobile and web browsers; (2) an

agreement with Mozilla to make Google the default search engine in the Mozilla web browser; and

(3) Android device agreements which make Google the default search engine on devices running the

Android operating system.

77. Apple. Since 2002, Google and Apple have been parties to an anticompetitive

exclusionary agreement, in which Google paid Apple an estimated $18 billion a year by 2021 to

make Google Search the default search engine in Safari, Apple’s web browser. Safari is the only

web browser that is pre-installed on Apple devices, including the iPhone, iPad, and Mac desktops.

48
Id. at 174.
49
Statista, Annual revenue of Google from 2002 to 2022, https://www.statista.com/statistics/266206/googles-annual-
global-revenue/#:~:text=In%20the%20most%20recently%20reported,billion%20U.S.%20dollars%20in%202022
(last accessed Oct. 30, 2023).

29
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 33 of 96

In exchange for default placement, Google currently pays Apple 36% of its revenue from search

traffic originating from the Safari browser search box.

78. As Whinston testified, this deal shows Google’s market power: “if Google—you

know, imagine if Google is involved in this—in negotiating this Apple contract, if it had equally

capable rivals, it wouldn’t be able to make that kind of money. Apple would simply them off

against each other. So, when you see that level of profit, it’s telling you that there’s a really big

gap and they have a lot of market power.” 50

79. This 2002 exclusionary agreement between Google and Apple ensured Google’s

ability to maintain its Search and Search Advertising monopolies. Testifying in the DC DOJ trial,

Google’s CEO conceded that such defaults are “very valuable.” 51 During his testimony, it also came

out that Apple was interested in modifying Google’s status during contract renewal discussions in

2007, but Google’s default positioning was too valuable. Google paid off Apple not to alter the terms

of the deal:

Apple had sought a choice screen in 2007, when Pichai was a Google
executive but years before his elevation to CEO. [DOJ attorney] Bellshaw
confronted Pichai with internal Google communications discussing an Apple
request to change the information services agreement that, to this day, makes
Google the default search engine on the Safari browser of every iPhone and
Mac sold in the United States.
During the 2007 round of contract renewal talks, according to the document,
Apple wanted to make Google one of two choices upon first use of the Safari
browser, all while maintaining financial terms under which it is paid billions
of dollars every year in shared revenue earned from advertising that
accompanies Google search queries. Pichai noted on the stand Monday that
Apple's request specifically covered a new version of Safari to be introduced
on Windows computers.
Google's internal discussions of the request, according to Bellshaw, helped

50
Trial Tr., United States, et al. v. Google LLC, No. 1:20cv03010 (D.D.C. Oct. 6, 2023) (Day 18) at 4775:7-13.

51
See Trial Tr., United States, et al. v. Google LLC, No. 1:20cv03010 (D.D.C. Oct. 30, 2023) (Day 30) at 7684:18-
20.

30
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 34 of 96

underscore the power of defaults that the company has continuously tried to
downplay during the antitrust bench trial now in its eighth week — Google
contends instead that users can easily switch away but choose not to. At the
time of Apple's request, according to internal communications, defaults
amounted to a “typically 75% take rate. Defaults have strong impact.” 52

80. The deal between Google and Apple was more than a mere business contract between

two parties. Jeff Shardell, the former Director of Business Development for Google, wrote a June 4,

2007, presentation for Pichai that described the deal as the “Apple Inc. Search Partnership.” 53 Google

and Apple had agreed to share the revenue that Google obtained by being given default status on

Apple devices. Tim Cook, the CEO of Apple, reportedly indicated to representatives of Google in

December of 2018 that “I imagine us as being able to be deep, deep partners; deeply connected where

our services end and yours begin and see[] no natural impediment to us working together.” 54

81. At one point during contract renewal discussions in 2016, Google thought Apple

desired to give end-users other suggestions for redirection. In a 2018 e-mail, Joan Braddi (“Braddi”),

Google’s head of product partnerships, said “[t]his concerned us which is why we added into the

agmt that they could not expand further than what they were doing in Sept 2016 (as we did not wish

for them to bleed off traffic). Also, they can only offer a ‘Siri’ suggestion exclusively for quality and

not because they want to drive traffic to Siri.” 55 The agreement nakedly restrained trade between

52
Bryan Koenig, Google CEO Admits Apple Deal To Be Default is ‘Valuable’, LAW 360 (Oct. 30, 2023),
https://www.law360.com/articles/1737914.The same article notes that Google applied a double standard here; while
espousing an exclusive default position for its search engine, in 2005, it told Microsoft that the latter should not take
any similar step with respect to an update of its operating system. Ex. No UPXO172 (1:20-cv-03010-APM),
https://www.justice.gov/d9/2023-10/417451_0.pdf (last accessed Oct. 30, 2023).
53
Ex. No. UPXO126 (1:20-cv-03010-APM), https://www.justice.gov/d9/2023-10/417441_0.pdf (last accessed Oct.
30, 2023).
54
Ex. No. UPX0617 (1:20-cv-03010-APM), https://www.justice.gov/d9/2023-10/417460.pdf (last accessed Nov. 8,
2023).
55
Ex. No. UPXO309 (1:20-cv-03010-APM), https://www.justice.gov/d9/2023-10/416999.pdf (last accessed Oct. 29,
2023).

31
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 35 of 96

Google and a potential competitor.

82. Obtaining this favored position was critical for Google. Professor Antonio Rangel, an

expert witness for the government in the DC DOJ Case, summarized key internal Google documents

between 2007 and 2017, which spoke of the “Power of Defaults.” 56

83. Google’s “pay to play” arrangement was disclosed publicly for the first time on

October 27, 2023 in the DC DOJ trial, during the testimony of Prabharkar Raghavan, one of Google’s

Senior Vice-Presidents. Google paid over $26.3 billion for such privileges in 2021 (with $18 billion

going to Apple alone) and $18.5 billion in 2020. 57 In 2020 and 2021, Google’s search advertising

56
Ex. No. UPXD101 (1:20-cv-03010-APM), https://www.justice.gov/d9/2023-09/416682.pdf (last accessed Oct. 29,
2023).
57
Bryan Koenig, Google Trial Reveals $26B Spent on Search Distribution, LAW360 (Oct. 27, 2023),
https://www.law360.com/competition/articles/1737865/google-trial-reveals-26b-spent-on-search-
distribution?spotlight=1 (“Koenig Article”); David Pierce, Google reportedly pays $18 billion a year to be Apple’s
default search engine, The Verge (Oct. 26, 2023), https://www.theverge.com/2023/10/26/23933206/google-apple-
search-deal-safari-18-billion.

32
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 36 of 96

revenues were $102.9 billion and $146.4 billion, respectively. Again, Google was sharing with Apple

a portion of its revenues obtained through Google’s default search engine status on Apple devices.

On November 7, 2023, during the trial in the DC DOJ Case, Jamie Rosenberg (“Rosenberg”), a

former Google executive, was cross-examined on this agreement and was asked whether, “[g]iven

the intensity of competition… ‘did you ever tell anyone [that] Google should not be paying billions

per year to Apple?’ Rosenberg responded that he doesn’t think so.” 58

84. According to testimony by Pichai in a separate trial a separate involving Google and

Epic Games (“the EPIC Games Case”), taking place in federal district court in the Northern District

of California, Google currently pays Apple 36% of the advertising revenue it earns from being the

default search engine on Apple’s Safari web browser. 59

85. Mozilla. In addition to the Apple foreclosure deal, Google entered into a similar

agreement with the browser provider Mozilla. Under that contract, Mozilla agreed to make Google

the default search engine for all search access points on its Firefox browser in exchange for a share

of search advertising revenue from searches originating from Firefox’s search box or the Firefox

homepage. Google has also secured agreements with smaller browser providers (Opera and UCWeb)

to be the default search engine in exchange for revenue-share payments.

86. These agreements effectively lock in Google as the dominant search engine, and

therefore the dominant search referral provider to Publishers, providing Google control over the

58
Matthew Perlman, Judge Told Google Helped Innovate Mobile Market, LAW360 (Nov. 8, 2023,
https://www.law360.com/competition/articles/1764723?nl_pk=787d704d-431c-432f-ba36-
94008c81ee47&utm_source=newsletter&utm_medium=email&utm_campaign=competition&utm_content=2023-
11-09&read_main=1&nlsidx=0&nlaidx=1. (“Perlman Article”).

59
The case is Epic Games Inc. v. Google LLC, et al., No. 3:20-cv-05671 (N.D. Cal.) (“Epic Games Case”). See
Bonnie Eslinger, Google CEO Denies App Store Monopoly in Epic Games Trial
LAW 360 (Oct. 30, 2023), https://www.law360.com/classaction/articles/1766669/-google-ceo-denies-app-store-
monopoly-in-epic-games-trial.

33
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 37 of 96

terms with which it deals with Publishers, because Publishers have no meaningful alternatives to or

substitutes for Google as a search intermediary.

87. Android Agreements. Google has also made exclusionary deals with: (1) original

equipment manufacturers of Android devices, like Samsung and (2) phone carriers that sell Android

devices, like Verizon. The Android operating system (“Android OS”) is a mobile phone operating

system that Google acquired in 2005. It is the second most widely used mobile operating system in

the U.S., behind Apple’s iOS. Android OS is open source, so numerous equipment manufacturers

can use Android OS on their mobile devices. Google has entered into two kinds of contracts with

original equipment manufacturers and phone carriers: Mobile Application Distribution Agreements

(“MADAs”) and Revenue Share Agreements (“RSAs”). 60 Pursuant to MADAs, original equipment

manufacturers are required to preinstall the Google Search app and Chrome browser and place the

app on the home screen. The RSA ensures that all preinstalled access points will have Google as the

preset default, with no competing search engine preinstalled.

88. Each of these types of agreements forecloses competition in the general search

services market for most Android phones and is inherently anticompetitive. For example, in 2020,

Google committed to paying Samsung (an original equipment manufacturer) $8 billion over four

years to secure Google’s search engine, voice assistant, and Play Store as the default on Samsung’s

mobile devices. Like its conduct with Apple and Mozilla, this conduct forecloses competing search

engines from gaining a foothold. This forces Publishers to deal with Google as an indispensable

provider of search referral traffic services on Google’s extortionate terms.

89. As a direct result of these foreclosure agreements, Google captures and traps users in

its search engine, thereby starving the market of meaningful competition for online search services.

60
See SJ Op. at 13-14, 40-41.

34
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 38 of 96

4. Tying.

90. To attract users to its walled garden search engine, Google ties the provision of its

general search engine services to the provision of services for digital news and reference publishing.

Under this anticompetitive tying arrangement, the tying market is online search services and thetied

market is the market for digital news and reference publishing services. 61 . Users who “purchase”

Google’s general search services through their attention and data, are forced to purchase Google’s

news and reference publishing services as well, because Google’s digital news publishing services

cannot be decoupled from search. On the supply side, Publishers who wish to purchase Google’s

search traffic referral services, must also purchase Google’s digital news publishing services. As

explained in the following section, this tying arrangement serves both to reinforce Google’s

monopoly power in the tying market of general search engine services, as well as expand Google’s

market power and harm competition in the tied market for digital news and reference publishing.

a. Google’s Extractive Relationship with Publishers.

91. The search engine transformed the traditional print media industry. Before the internet

became ubiquitous, Americans obtained original, fact-based informational content from newspapers,

magazines, and reference works (back issues, encyclopedias, dictionaries, almanacs, etc.). Since the

emergence of the web, the news and reference industry has shifted towards online distribution and

consumption, as part of the digital content media ecosystem. The percentage of Americans getting

at least some of their news online rose from 2% in 1995 to 93% in 2018. 62 And, today, only one

publisher of encyclopedias—World Book—remains in print. Digitization has led to the development

Tying also occurs with respect to Google’s search advertising services and its implementation of the recent
61

modification of its AdSense billing practices, as discussed in a separate section below.

62
News Media Alliance, Google Benefit from News Content: Economic Study at 3 (June 2019),
https://www.newsmediaalliance.org/wp-content/uploads/2019/06/Google-Benefit-from-News-Content.pdf.

35
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 39 of 96

of new formats for publishing informational content online, including digital magazines and

newspapers (e.g., The Helena World News, Washington Post) as well as blogs and other
63
informational websites (e.g., AARP, AccuWeather, Consumer Reports, WebMD). Digital

publishing is a substantial line of commerce. In the United States, digital publishing revenue is

expected to exceed $27 billion by 2027. 64

92. The search engine introduced a new way to distribute news products. In the analog

era, readers got the news from brick-and-mortar retailers, paper boxes, and “delivery boys.” Then,

in the late 1990s, as publishers began issuing digital versions of their products, the first generation

of search engines—Yahoo!, AltaVista, Ask Jeeves, and Google—provided publishers a unique

intermediation service. Search engines can connect users seeking information to publishers

providing answers, by generating a list of search results linking to their relevant webpages. The flow

of users from the search engine to publishers is called search-referral traffic.

93. After a decade on the market, Google eclipsed its rival search engines, reaching an

82.5% market share in November 2009. 65 Because Google monopolizes search, it exercises

dominance over Publishers, who play a dual role in the multi-sided general search services market:

they are both purchasers of search-referral traffic and suppliers of content.

94. Google’s commercial relationship with Publishers has unique market dynamics. Both

Google and Publishers have the same triple-product business model: (1) content attracts users; (2)

63
In 2001, in recognition of this distinct emerging market, a trade association called the Online Publishers Association
(now called Digital Content Next) was formed. See https://digitalcontentnext.org/.

Statista, U.S. digital publishing industry—statistics and facts (Aug. 30, 2023),
64

https://www.statista.com/topics/1453/digital-publishing/#topicOverview
65
Statcounter: GlobalStats, Search Engine Market Share United States of America,
https://gs.statcounter.com/search-engine-market-share/all/united-states-of-america/2009 (last accessed Oct. 29,
2023).

36
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 40 of 96

users provide their attention and data; and (3) attention and data are used to sell advertising and

improve services.

95. Google and Publishers exchange the inputs to this model: content provided by

Publishers and referral traffic (i.e., user attention and data) provided by Google.

96. This exchange occurs in billions of transactions as Google serves search results in

response to user queries seeking information. In a competitive search market, this relationship

would be symbiotic, with the value generated by the exchange being split fairly and sustainably.

But in a monopolized search market, the relationship turns parasitic. The monopolist—Google—

can extract rent profits by raising the cost of referral traffic (by extracting more content) and

depressing the quality of the referral-traffic service (by diverting traffic).

97. Publishers have no choice but to accept this parasitic arrangement because they are

locked in by Google’s gatekeeping role as a dominant digital platform. Google locks them in by

leveraging barriers to entry and technical features of the search engine.

98. Searching the web is a three-step process requiring high investments and massive

computing power. Google Search involves these stages:

37
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 41 of 96

• Crawling: Google downloads text, images, and videos from pages it found
on the internet with automated programs called crawlers.
• Indexing: Google analyzes the text, images, and video files on the page,
and stores the information in the Google index, which is a large database.
• Serving search results: When a user searches on Google, Google returns
information that's relevant to the user's query. 66

99. Google uses “a huge set of computers to crawl billions of pages on the web” using a

web crawler program called “Googlebot.” 67 Googlebot renders and copies the content of the

webpage. Google uses that ingested content to index the webpage, making the webpage searchable

and findable through Google Search. The relevance of search results is determined by applying an

algorithm to pull a responsive list of webpages from the index.

100. “A search engine can function only if it has access to an index, and an index can exist

only once web pages have been crawled and collated into a repository.” 68 This requires high fixed

costs and sizable server storage and computing power. These barriers have concentrated English-

language search indexing into the hands of the only two companies that maintain comprehensive

indexes of the web: Google and Bing (Microsoft). 69 “Other search engines—including Yahoo and

DuckDuckGo—must purchase access to the index from Google and/or Bing through syndication

agreements that provide syndicated search engines with access to search results and search

advertising.” 70

101. Google’s gatekeeping role in web crawling and indexing enables it to lock in

publishers due to switching costs. Crawlers slow down websites and can cause them to crash, so

66
Google, In-depth guide to how Google Search works,
https://developers.google.com/search/docs/fundamentals/how-search-works (last accessed Oct. 29, 2023).

67
Id.

68
House Subcommittee Report at 78.

69
Id. at 80.

70
Id.

38
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 42 of 96

most website publishers block access to all but a few crawlers. 71 As the House Subcommittee Report

explained:

The one crawler that nearly all webpages will allow is Google’s “Googlebot,” as
disappearing from Google’s index would lead most webpages to suffer dramatic drops
in traffic and revenue. Any new search engine crawler, by contrast, would likely be
blocked by major webpage owners unless that search engine was driving significant
traffic to webpages—which a search engine cannot do until it has crawled enough
webpages. 72

102. The next section will explain how Google has leveraged its monopoly power in

search to misappropriate Publishers’ content and force them to supply an unlawful tying

arrangement that makes them compete against their own products.

b. Building the Walled Garden: Google Misappropriates

Publisher Content

103. Google began life as an information retrieval portal, designed to help users find

links to relevant webpages and navigate away from Google as fast as possible. This is what users

wanted from a search engine. And Google’s original PageRank algorithm gave consumers what

they wanted. As Google’s co-founder Larry Page told an interviewer in 2004: “We want to get you

out of Google and to the right place as fast as possible.” 73

104. But Google quicky realized that this goal was at odds with Google’s core business

plan: selling ads against search. In 2000, Google launched the AdWords program, selling space on

Google’s SERP. Advertisers could bid for keywords and their ads would be targeted to users as

sponsored links appearing alongside (and later above) the “organic” search results ranked by their

71
Hal Singer, Written Comments in Response to U.S. Copyright Office’s Publishers’ Protection Study: Notice and
Request for Public Comment, 86 Fed. Reg. 56721 (Oct. 12, 2021),
https://www.copyright.gov/policy/publishersprotections/initial-comments/Hal%20Singer%20-
%20Initial%20Comment.pdf

72
Id. at 79.

73
David Sheff, Playboy Interview: Google guys: a candid conversation with America’s newest billionaires about
their oddball company, how they tamed the web and why their motto is “Don’t be evil”, Playboy, 55 (Sept. 1, 2004).

39
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 43 of 96

relevance to the user’s query. Today, search advertising generates the bulk of Google’s revenue,

accounting for roughly 79%. 74

105. By 2010, Google faced several problems. First, the bulk of Google searches (an

estimated 80%) are for non-commercial information—i.e., timely news or evergreen reference

content. But Google does not serve ads for these searches. To monetize these searches, Google

needed to entice information seekers to stay engaged with the search platform and use Google for

commercial queries as well.

106. Second, advertising degraded the quality of Google’s product: search. As one Wall

Street Journal reporter put it in 2011: “Speaking as a consumer, I find my Google search results to

be more and more polluted with junk that I don’t want to see or that doesn’t seem relevant.” 75

107. To overcome these problems, Google needed to attract users to Google Search with

high-value informational content and “trap” them within Google’s ecosystem.

108. Beginning in or around 2010 or 2011, Google devised a long-term plan to

transform Google into a walled garden. Since then, and throughout the Class Period, Google has

maintained its monopoly in search by transforming itself from an information retrieval service to a

question-and-answer service, aimed at displacing other digital publishers.

109. In May of 2011, Google CEO Eric Schmidt explained this strategy:

we’re trying to move from answers that are link-based to answers that are
algorithmically based, where we can actually compute the right answer. And we now
have enough artificial intelligence technology and enough scale and so forth that we can,

74
Statista, Net Search Advertising Revenue of Google in the United States from 2019 to 2024,
https://www.statista.com/statistics/271527/forecast-of-revenues-from-paid-search-in-the-us/ (last accessed Oct. 30,
2023).

75
Joshua Benton, Google now wants to answer your questions without links and with AI. Where does that leave
publishers?, Nieman Lab (Feb. 7, 2023), https://www.niemanlab.org/2023/02/google-now-wants-to-answer-your-
questions-without-links-and-with-ai-where-does-that-leave-publishers/

40
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 44 of 96

for example, give you — literally compute the right answer. 76

110. To publish answers to users’ questions, Google needed a supply of information.

Normally, a publisher obtains their inventory in two ways: (1) produce it through original reporting

and research or (2) syndicate it through licensing arrangements with other publishers. But Google’s

monopoly power in search gave it a third option: misappropriate content from Publishers.

111. Google already had free access to high-quality news and reference content thanks

to its general search services. To be discoverable in Google’s search results, and receive Google’s

search-referral traffic services, must allow Google to crawl their websites, copy their original

content, and make vast and unknown uses of that content for Google’s benefit. As a practical matter,

Publishers can block Googe’s web-crawler, Googlebot. They can do so by embedding a robots.txt

file or a “no index” rule in the code of their webpage. But this would “drop that page entirely from

Google Search results.” 77

112. For Publishers operating on thin profit margins, going dark on Google would be

commercial suicide. Search referral traffic is the largest single source of external traffic to news

websites: exceeding social media such as Facebook or Twitter. 78 Google controls the overwhelming

share of search-referral traffic to Publishers: 95% of all referral traffic to news websites from search

engines originates from Google, only 5% originates from a handful of competitors: Bing, Yahoo,

and DuckDuckGo. 79

76
Joshua Benton, Eric Schmidt: Google wants to get so smart it can answer your questions without having to link
you elsewhere, Nieman Lab (June 1, 2011), https://www.niemanlab.org/2011/06/eric-schmidt-google-wants-to-get-
so-smart-it-can-answer-your-questions-without-having-to-link-you-elsewhere/.

77
Google, Block Search indexing with no index, https://developers.google.com/search/docs/crawling-
indexing/block-indexing (last accessed Oct. 30, 2023).

78
Press Gazette Article.

79
Id.

41
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 45 of 96

113. Informational search (i.e., news and reference search) constitutes a line of commerce

(or alternatively a market) in its own right. Google treats news and reference search as its own

vertical sub-set, bundled within its general search services: Google’s algorithm detects an

informational request and serves specialized content and links on the SERP. Because Google

controls 95% of such referral traffic, it can extract an exorbitant price from Publishers: unlicensed

use of their product inventory.

c. Google Ties General Searches to News and Reference

Publishing.

114. In 2012, Google began publishing news and reference content directly on the SERP,

providing instant answers to user queries. Launched in May 2012, the Knowledge Graph was

Google’s first foray into replacing search result links with rich-text answers. When a user searches

for information on a topic, Google displays a “Knowledge Panel” to the right of the search results.

This panel contains a summary of content drawn from the Knowledge Graph database. Google

compiled this massive database by extracting information from Publishers’ websites—what

Google calls “materials shared across the web”—and from “open source and licensed databases.”

By 2020, the Knowledge Graph had grown to “500 billion facts about five billion entities.”80

Google described the Knowledge Graph as a “critical first step towards building the next

generation of search, which taps into the collective intelligence of the web and understands the

world a bit more like people do.” 81 Much of the “collective intelligence” Google tapped into was

content misappropriated from Publishers.

80
Danny Sullivan, A reintroduction to our Knowledge Graph and knowledge panels, Google The Keyword (May
20, 2020), https://blog.google/products/search/about-knowledge-graph-and-knowledge-panels/.

81
Amit Singhai, Introducing the Knowledge Graph; things, not strings, Google The Keyword (May 16, 2012),
https://blog.google/products/search/introducing-knowledge-graph-things-not/.

42
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 46 of 96

115. That same year, Google launched a project to transform Google Search into an

information-publishing platform using AI. Google put two engineers, Steven Baker and Srinivasan

Venkatachary, in charge of a project focused on “Question Answering from the Web.” 82 Internally,

the project was called “WebAnswers.” 83 Google merged the Knowledge Graph question-

answering efforts into Baker and Venkatachary’s team, with a goal to develop “one coherent

question answering system” using “web extraction.” 84

116. In 2016, “Featured Snippets” was launched. When a user asks a question in Google

Search, Google algorithmically generates an answer by extracting a summary from a webpage and

displaying it in an information “box” on top of the search results. Hardly a snippet, the summary

can contain entire paragraphs.

117. Google’s patent “Natural Language Results for Intent Queries” (2016) makes clear

that the goal of Featured Snippets is to shift users away from clicking on links to news and

reference sources by publishing extracts of such content directly on the SERP. 85 The patent

explains that Google’s traditional search results “fail to provide a complete, easily understood

answer non- factual questions where there is no one correct answer.” 86 From Google’s perspective,

the problem with its hyperlinked search results is that they make the user navigate away from

Google’s SERP: “While a user can select the link associated with the snippet to view the context

Srinivasan Venkatachary, LinkedIn Profile, https://www.linkedin.com/in/srinivasanvenkatachary (last accessed

82

Oct. 30, 2023).

83
Steven Baker, LinkedIn Profile, https://www.linkedin.com/in/steven-baker-5077885/ (last accessed Oct. 30,
2023).

84
Id.

85
US Pat. No. 9,448,992 B2 to Shmiel et al., Sep. 20, 2016,
https://patentimages.storage.googleapis.com/04/44/5d/1acc228f2d34c3/US9448992.pdf.

86
Id. at col. 1: 6-22.

43
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 47 of 96

of the snippet in the original document to determine whether the identified information is adequate,

this slows the user experience and involves additional effort on the part of the user to receive an

answer to a non-factual question.” 87 Google’s solution is to provide “[n]atural language answers .

. . in a paragraph and/or list format that provide diverse or complex answers or more than one fact

per answer.” 88

118. Google admits that its natural-language answers have value because they are

extracted from news and reference content. As the patent states: “The natural language answers

are of high quality because they are derived from authoritative sources.” 89

119. Google’s “Featured Snippets” are effective because they extract the most valuable

part of publishers’ inventory: the “lead.” This is the first paragraph that presents the main points of

an article, using journalism’s classic inverted-pyramid structure.

Figure 1: “Inverted pyramid in comprehensive form” by Christopher Schwartz is licensed

under CC BY-SA 3.0

87
Id.

88
Id. at col. 4: 3-5.

89
Id. at col. 4: 5-7.

44
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 48 of 96

120. Google extended Featured Snippets through the “People Also Ask” box—a SERP

feature that displays a drop-down list of follow-up questions related to the user’s original search

query. When a user clicks on a suggested question, Google displays yet another natural-language

answer. With each click, more questions appear below it, followed by more answers. A user can

view hundreds of questions and view hundreds of answers—extracted from Publishers—all

without ever leaving Google’s SERP.

121. The Featured Snippets (WebAnswers) system has transformed Google’s SERP

from a list of simple blue hyperlinks to something closer to front-page news or an encyclopedia

entry, with rich-text content and high-quality photography. The following figures illustrate this

evolution. 90

90
Vision Museum, https://developers.google.com/search/docs/crawling-indexing/block-indexing (last accessed Oct.
30, 2023).

45
Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 49 of 96

46
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 50 of 96

122. In 2023, Google Search is now a digital publishing platform that competes with other

Publishers for attention, user engagement, and advertising revenue. 91 When Google “detect[s] a

search query is news-oriented” it serves a SERP that provides the news itself. 92

123. For example, “what is happening with Sam Altman?” returns a SERP dominated by a

91
Google Search is in fact just one part of Google’s walled garden for news. Google admits that it has a growing suite
of news publishing products including: Google News (a news aggregator website and app), Discovery (a news feed
extension for search), YouTube, and Google’s Voice Assistant. See Google, Presenting news results in helpful ways,
https://newsinitiative.withgoogle.com/hownewsworks/approach/presenting-news-in-helpful-ways/ (last accessed
Oct. 30, 2023).

92
Id.

47
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 51 of 96

full 48-word paragraph of journalism copied from PBS: “Sam Altman is back as the chief executive

of OpenAI. The hot tech start-up behind ChatGPT not only brought Altman back; it’s also

overhauling the board that fired him with new directors, ending a dramatic five-day standoff that's

transfixed Silicon Valley and the artificial intelligence industry.”

124. Not only does this SERP fully answer the user’s question, but the “Videos” box

that appears below the “Featured Snippet” only returns links to YouTube—Google’s video

platform. Google has achieved a self-preferencing walled garden. Other organic links from

48
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 52 of 96

Publishers are pushed down “below the fold,” 93 meaning a user would have to scroll down to

actually find links to news websites. But users have no reason to click-through, because they have

consumed the news directly on Google Search.

d. Google’s Scheme Maintains its Monopoly in the Tying Search

Market and Harms Competition in the Tied Publishing
Market.

125. Google’s strategy is an anticompetitive tying arrangement. Throughout the Class

Period, Google has tied its search engine product to misappropriated news and reference content

published on its search platform. The tying product market is the general search services market,

where Google exercises enduring monopoly power. The tied product market is digital news and

reference publishing. When a user searches for information on Google, buying Google’s search

service is conditioned on also buying Google’s re-published news and reference content.

126. This tying arrangement has been enormously successful for Google. Informational

news and reference content is valuable for Google because it is in high demand by Google users.

Trial evidence in the DC DOJ Case reveals that roughly 80% of all searches on Google are non-

commercial, informational queries, according to Google’s expert witness, Mark Israel. 94 Google

does not serve search ads for these searches. Instead, it monetizes them as “free samples for

advertising” to “build its brand.” 95

127. Long ago, Google admitted publicly that it uses news content to attract and trap users.

In 2008, Marissa Mayer, then Google’s Vice-President for Search Products and User Experience,

clarified that news is valuable because it drives search traffic. According to a Forbes conference

In print publishing, “below the fold” refers to articles appearing below the horizontal fold of a newspaper, which
93

were less likely to garner attention.

94
Trial Tr., United States, et al. v. Google LLC, No. 1:20cv03010 (D.D.C. Nov. 2, 2023) (Day 33) at 8608:13-
8609:23.

95
Id.

49
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 53 of 96

attendee, Mayer explained that Google News—a news aggregation product without ads—“funnels

readers over to the main Google search engine, where they do searches that do produce ads. And

that’s a nice business. Think of Google News as a $100 million search referral machine.” 96

128. Google’s tying arrangement is highly effective at diverting users into Google’s

walled garden and reducing referral traffic to rival Publishers. A 2017 study analyzed two million

Featured Snippets and found that they cause a significant drop in the click-through rate to websites

appearing in the organic search results. 97

129. Google’s tying arrangement has produced a zero-click world. As of 2020, roughly

65% of all Google searches were zero-click searches, where users get the information they seek on

the search results page rather than clicking the link to the publishers’ content. 98 That same year, a

staggering 77% of Google searches on mobile phones were zero-click searches. 99 All of this search

traffic bolsters the scale and network effects that protect and maintain Google’s monopoly in general

search services.

130. By diverting information-seekers to Google Search, Google’s tying arrangement has

made Google.com the largest news and reference website in the world. In 2023, Google.com has

96
Jon Fortt, What’s Google News worth? $100 million, Fortune (July 27, 2008),
https://fortune.com/2008/07/22/whats-google-news-worth-100-million/.

97
See Tim Soulo, Ahrefs’ Study of 2 Million Featured Snippets: 10 Important Takeaways, Ahref’s Blog (Apr. 7,
2020), https://ahrefs.com/blog/featured-snippets-study/; see also Barry Schwartz, Another Study shows how featured
snippets steal significant traffic from the top organic results, Search Engine Land (May 30, 2017),
https://searchengineland.com/another-featured-snippet-study-shows-steal-significanttraffic-first-organic-result-
275967 (summarizing Ahrefs’ study).

98
Rand Fishkin, In 2020, Two Thirds of Google Searches Ended Without a Click, Sparktoro (Mar. 22, 2021)
(“Fishkin Article”), https://sparktoro.com/blog/in-2020-two-thirds-of-google-searches-ended-without-a-click/.

99
Id.

50
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 54 of 96

approximately 87 billion visits per month. 100 Roughly eighty percent of those visits are

informational, which means that 69.6 billion of Google’s monthly visits are by consumers of

informational digital content. Many of these visitors specifically seek conventional news content: in

the second quarter of 2023, “weather” was the second and “news” was the ninth most common search

query on Google. 101 With an audience of 69.6 billion information consumers per month, Google.com

is far and away the largest publisher of online news and reference content. Google produces none of

this content: instead, it exacts unlicensed syndication from Publishers’ inventories copied by

Google’s web-crawler.

Global Monthly Visits to News & Media Publishers

October 2023

google.com - (informational queries) 69.6B wikipedia.org - 4.54B

yahoo.com - 3.58B cnn.com - 707.3M
msn.com - 665.5M nytimes.com - 642.1M
foxnews.com - 331.7M news.yahoo.com - 251.3M
nypost.com - 158.8M washingtonpost.com - 155.4M

131. Google’s tying arrangement substantially lessens competition in the tied market of

digital news and reference publishing by free-riding and raising rivals’ costs. As a provider of search-

referral traffic services to Publishers, Google is raising the cost of its services—by coercing

Similarweb.com, google, https://www.similarweb.com/website/google.com/#ranking (last accessed Oct. 30,

100

2023).

101
Statista, Top Google search inquiries worldwide in the 2d quarter of 2023,
https://www.statista.com/statistics/265825/number-of-searches-worldwide/2023 (last accessed Oct. 30, 2023).

51
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 55 of 96

unlicensed use of web-crawled content. At the same time, Google is reducing the quality of its

service by reducing the volume of referral traffic.

132. Google claims it sends users “to news sites 24 billion times per month,” underscoring

that it is an essential channel of news dissemination. 102 But this figure only represents users who

actually click on links to news sites. Since 69% of Google searches are zero-click searches, tens of

billions of users are consuming the news directly on Google Search. Yet Google pays none of the

labor and overhead costs that make reporting the news possible. Every user who consumes news and

reference information on Google is a stolen customer for the original Publisher of that content.

Publishers cannot monetize that customer through display advertising or subscriptions.

133. Publishers have no choice but to supply content to a tying arrangement that is forcing

them to compete against their own products. Google does not permit Publishers to fully block

Featured Snippets without blocking all snippets, which would cause search referral traffic to

plummet. 103 As the News Media Alliance notes, this “would be suicidal.” 104

134. By imposing monopoly rents on Publishers, Google is raising their costs to levels that

often cannot sustain production. Google’s anticompetitive conduct is decimating news Publishers’

revenue which has fallen from roughly $50 billion in 2005 to $20 billion in 2022. Like a parasite

killing its host, Google is forcing smaller publishers out of business and forcing even larger

publishers to shrink staff and operations. Since 2005, America has lost more than a quarter of its

102
Richard Gingras, Setting the record straight on news, Google The Keyword (June 26, 2020),
https://blog.google/outreach-initiatives/google-news-initiative/setting-record-straight-news/.

103
See Google, Featured snippets and your website, https://developers.google.com/search/docs/appearance/featured-
snippets (“Using a low max-snippet setting doesn't guarantee that Google will stop showing featured snippets for
your page. If you need a guaranteed solution, use the nosnippet rule.”) (emphases in original).

104
News Media Alliance, How Google Abuses Its Position as a Market Dominant Platform to
Strong- Arm News Publishers and Hurt Journalism at 23 (Updated: September 2022)
https://www.newsmediaalliance.org/wp-content/uploads/2022/09/NMA-White-Paper_REVISED-Sept-2022.pdf.

52
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 56 of 96

newspapers (2,500) and is expected to lose a third by 2025. 105 Seventy million Americans now live

in news deserts, with little to no local news coverage, according to Northwestern University’s 2022

State of Local News report. 106 Google is contributing to the decline of the news industry.

135. Google’s tying arrangement distorts consumers’ choices. But for the tie, consumers

would choose to acquire content from the Publishers who actually produce it and browse a variety

of publications. Instead of a zero-click world, there would be a click-through world, as existed

before Google implemented the tie. This would help sustain production, which ultimately benefits

consumers. Instead, by squeezing Publishers out of the market, Google’s tying arrangement

reduces the variety of news and reference products on the market.

136. It also reduces the quality of news and reference content on the digital market. First,

Google’s tying arrangement degrades the quality of news and reference information by publishing

only the lead, while stripping away nuance and context. As the Washington Post noted:

Google’s knowledge panels regularly, if inadvertently, make rather important decisions for
us: Taiwan, you’ll remember, is described as if it were an independent nation, when only
22 countries actually recognize it as such. Meanwhile, Google corrects searches for
“Londonderry,” Ireland’s fourth-largest city, to “Derry,” the (unofficial) term favored by
Irish nationalists. 107

Google’s rich-text answers are an inferior product to the full news and reference articles they compete

against. But because they are tied to search, they proliferate in the marketplace of ideas, as substitutes for

original content.

137. Second, many of Google’s Featured Snippets and Knowledge Boxes display content

Penny Abernathy, The State of Local News – The 2022 Report, Northwestern Local News Initiative (June 29,
105

2022), https://localnewsinitiative.northwestern.edu/research/state-of-local-news/report/.

106
Id.

107
Caitlin Dewey, Google’s sketchy quest to control the world’s knowledge, Washington Post, May 11, 2016,
https://www.washingtonpost.com/news/the-intersect/wp/2016/05/11/you-probably-havent-even-noticed-googles-
sketchy-quest-to-control-the-worlds-knowledge/.

53
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 57 of 96

extracted from potentially unreliable sources of information that are not professionally edited. For

example, as a 2016 Washington Post article observed, Google’s knowledge panels can extract

information “from unsourced Wikipedia edit[s].” 108

138. Third, Google’s tied rich-text answers distort consumer choice because they

“algorithmically confer a lot of unearned authority.” 109 Information displayed by Google is

“‘atomized’ or removed from its source and placed alongside other content.” 110 This makes it

harder for consumers to choose between products or built brand trust based on informed judgments

on editorial quality. By divorcing content from its professional producers, Google makes it harder

for consumers to distinguish between, on one hand, fabrication and opinion and, on the other,

trustworthy sources, produced with professional ethical standards.

139. By squeezing out rival Publishers, Google is contributing to the loss of tens of

thousands of reporters and editors from the labor market. Employment has dropped 70% from

2005-2022. With fewer reporters, less news is gathered on current events. And with fewer editors

and fact checkers, less online information is verified for accuracy under the ethics and standards

of professional journalism.

140. As a result, misinformation proliferates. Google does not merely pass on fake news

or inaccurate information. In the case of Bard, its new chatbot, it creates false information, as

discussed in paragraphs 151-52, in the next section of this complaint.

108
. Id.

109
Id.

110
House Subcommittee Report at 52.

54
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 58 of 96

5. Introduction And Implementation of Bard.

141. The tying conduct of Google described in the previous sections was only extended

and exacerbated by Google’s introduction of Bard in 2023.

142. In order to understand the most recent anticompetitive conduct in which Google engaged

as to Publishers, it is necessary to discuss the introduction of the form of GAI known as OpenAI in 2022

and Google’s response to it in 2023. OpenAI has operated through, inter alia, the programs ChatGPT-

3 (introduced in 2020), ChatGPT-3.5 (introduced in December of 2022), and ChatGPT-4

(introduced in March of 2023). 111

143. Microsoft was an early supporter of OpenAI, investing $1 billion in the company in

2019. 112 Its cumulative investment is $13 billion (including $10 billion in early 2023). As a result

of this latter investment, Microsoft reportedly has a 49% stake in the company and a right to 75%

of OpenAI’s profits until its investment is recouped.

144. The two companies describe themselves as having a “partnership.” 113 Indeed, going

back to 2020, Microsoft had given OpenAI access to its Azure infrastructure, a cloud platform for

more than 200 products and various services. 114 It was through this platform that OpenAI

disseminated or sold the ChatGPT-3.5, ChatGPT-4, and the ChatGPT-4 Plus applications.
Furthermore, as a result of this “partnership”, Microsoft became the exclusive cloud partner for

OpenAI, powering all workloads across products, Application Programming Interfaces, and

111
A timeline of corporate events for OpenAI is set forth in Sarah O’Neill, The History of OpenAI, LAX Hub (May
2, 2023), https://www.lxahub.com/stories/the-history-of-
openai#:~:text=Founded%20in%202015%20by%20a,Greg%20Brockman%2C%20and%20Andrej%20Karpathy
(“OpenAI History”).

112
Microsoft, OpenAI And Microsoft Extend Partnership, Official Microsoft Corporate Blog (Jan. 23, 2023),
https://blogs.microsoft.com/blog/2023/01/23/microsoftandopenaiextendpartnership/ (“Microsoft PR”).

113
Id.
114
Id.

55
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 59 of 96

resources. Microsoft integrated ChatGPT-4 into its Bing web browser and plans to add it to apps

like Word, PowerPoint and Outlook. 115 This feature was originally referred to as “Bing Chat” but

is now referred to as “Copilot.”

145. Google was threatened by this new competitive element from its biggest rival.

Although Google initially dismissed the idea of introducing its own GAI, saying it had a greater

“reputational risk” than a startup like OpenAI,116 that attitude changed rapidly, given the

groundswell of public attention over ChatGPT. By January of 2023, Google announced the

introduction of Bard and other GAI programs. As one source stated:

Following its much-reported management declaration of a ‘Code

Red' response to Microsoft's repeated and increasingly more
substantial OpenAI investments and the startup's creation,
ChatGPT, it has been widely presumed that this latest
announcement is part of a strategic commercial countermeasure.
According to the New York Times, Google's founders, Larry Page
and Sergei Brin, were brought in for a significant C-suite meeting
to discuss the threat of ChatGPT to the ubiquitous search engine's
hold on the internet. 117

146. Bard had been in the planning and testing stage at Google for years. It was initially

based on the LaMDA Large Language Model (“LLM”), which had been trained on a dataset

consisting of: (a) 12.5% code documents from sites relating to programming, such as Q&A sites,

tutorials, etc.; (b) 12.5% Wikipedia (in English); (c) 6.25% English web documents; (d) 6.25% non-

115
Tom Warren, Microsoft to demo its new ChatGPT-like AI in Word, PowerPoint, and Outlook soon, The Verge
(Feb. 10, 2023), https://www.theverge.com/2023/2/10/23593980/microsoft-bing-chatgpt-ai-teams-outlook-
integration?uuid=FtJuUSu2JEtqP3Qp0716.
116
Britney Nguyen, Google execs say the company isn’t launching a ChatGPT competitor because if has greater
‘reputational risk’ than startups like OpenAI, Insider (Dec. 14, 2022), https://www.businessinsider.com/google-isnt-
launching-chatgpt-competitor-due-to-reputational-risk-2022-12.

117
Jonny Wills, Google in ‘Code Red’ response to ChatGPT, UC Today (Jan. 26, 2023),
https://www.uctoday.com/unified-communications/google-in-code-red-ai-response-to-chatgpt/.

56
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 60 of 96

English web documents; and (e) 50% “dialogs data from public forums.” 118

147. By the time Bard was introduced, it was based on PaLM2, Google’s latest data training

model, which has multilinguistic capabilities. Google is vague about the sources of training for

PaLM2, merely noting that “the system’s training corpus is comprised of ‘a diverse set of sources:

web documents, books, code, mathematics, and conversational data,’ without offering further

detail.”119

148. Bard, like ChatGPT, was trained on datasets that included news, magazine and

digital publications. A 2023 report from the News Media Alliance (“NMA”) states the following:

• In fact, our analysis of a representative sample of news, magazine,

and digital media publications shows that the popular curated datasets
underlying some of the most widely used LLMs significantly
overweight publisher content by a factor ranging from over 5 to
almost 100 as compared to the generic collection of content that the
well-known entity Common Crawl has scraped from the web.

• Other studies show that news and digital media ranks third among
all categories of sources in Google's C4 training set, which was used
to develop Google's GAI-powered search capabilities and products
like Bard. Half of the top ten sites represented in the training set are
news outlets.

• The LLMs also copy and use publisher content in generating

outputs. The LLMs can reproduce the content on which they were
trained, demonstrating that the models retain and can memorize the
expressive content of the training works. 120

149. Indeed, in 2023, it was publicly revealed that Google’s misappropriation of

118
Roger Montti, Google Bard AI—What Sites Were Used To Train It?, Search Engine Journal (Feb. 10, 2023),
https://www.searchenginejournal.com/google-bard-training-data/478941/#close.

James Vincent, Google announces PaLM2 language model, already powering 25 google services, The Verge
119

(Mar. 10, 2023), https://www.theverge.com/2023/5/10/23718046/google-ai-palm-2-language-model-bard-io.

120
News Media Alliance, White Paper: How the pervasive copying of expressive works ti train and fuel generative
artificial intelligence systems is copyright infringement and not a fair use at 1-2 (Oct. 20, 2023),
https://www.newsmediaalliance.org/wp-content/uploads/2023/10/AI-White-Paper-with-Technical-Analysis.pdf)
(“NMA White Paper”).

57
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 61 of 96

Publishers’ content was central to its development of GAI. Beginning in or around 2017, and

continuing throughout the Class Period, Google has “trained” its LLMs using millions of unlicensed

copies of Publishers’ content—including Plaintiff’s.

150. As noted above, Google was able to acquire massive amounts of Publishers’ content

because of its monopoly power in search. One of Plaintiff ‘s newspapers was one of the Publishers

subjected to this misappropriation. An April 2023 Washington Post investigation revealed that

Google had used copious amounts of Plaintiff’s content to train its LLM Bard (discussed below), as

part of “Google’s C4 data set, a massive snapshot of the contents of 15 million websites that have

been used to instruct some high-profile English-language AIs, called large language models,

including Google’s T5 and Facebook’s LLaMA.” 121 Google’s LLMs were trained on 650,000

“tokens” of text extracted from the The Helena World newspaper (helena-arkansas.com). 122

Kevin Schaul, et al, Inside the secret list of websites that make AI like ChatGPT sound smart, Washington Post
121

(Apr. 19, 2023), https://www.washingtonpost.com/technology/interactive/2023/ai-chatbot-learning/

122
“In tokenization, AI algorithms, models, or datasets are represented as tradable tokens on a blockchain. As a
result, developers may easily share, sell, or license their AI inventions to others within a decentralized ecosystem by
tokenizing them. Applications for tokenized AI models include data analysis, picture recognition, natural language
processing, and more.” Alisha Bains, What Are AI Tokens? A Comprehensive Guide, CCN (Aug. 29, 2023),
https://www.ccn.com/education/what-are-ai-tokens-a-comprehensive-guide/.

58
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 62 of 96

151. All of this web scraping helped Google enormously in developing search algorithms,

the complex mechanisms that retrieve information responsive to an end-user’s search inquiry. In

creating such algorithms, Bard is then able to rely on the data obtained from such scraping, the

indexing of that data, and the ranking of that data responsive to an inquiry. 123 LLMs are used to

decipher the content of the data and the end-user’s search request. The LLM is responsible for

ensuring that: (a) the search retrieves germane data, meaning that the quality of the website is deemed

responsive; (b) the usability of that website; and (c) the context of search history related to the

website.124 Thus, the data obtained from various sources, particularly news sources, is critical to

Google’s search capability, but the entities from which Bard’s trainers scrape data never authorize

123
Ben Lutkevich, Google algorithms explained: Everything you need to know, Tech Target (Apr. 20, 2023),
https://www.techtarget.com/whatis/feature/Google-algorithms-explained-Everything-you-need-to-
know#:~:text=What%20are%20Google%20search%20algorithms,keywords%20that%20match%20the%20query.

124
Id.

59
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 63 of 96

such web crawling and Google never shares any of its revenue from such scraping with them.

152. At the time of Bard’s introduction in February of 2023, many at Google believed that

the introduction was too rushed and that Bard was not yet ready for prime time. As explained in one

article:

Google’s artificial intelligence chatbot, also known as Bard AI has

been a fan favourite among regular people and AI enthusiasts as it
poses some competition to OpenAI’s ChatGPT. However,
Google’s employees are evidently not happy with the chatbot and
have been against how it was developed and launched.
Google’s employees have described BardAI as a quickly produced,
badly executed, and uncharacteristically Google-like product that
will do more harm than good for the company’s brand.
According to several reports, several engineering and non-
engineering staff who tested the chatbot referred to it as “a
pathological liar” and pleaded with the firm not to launch it.
The issue was raised during a discussion with eighteen current and
former Google employees. During one of these company-wide
sessions, an employee mentioned how frequently Bard would give
users potentially dangerous advice, whether it was about how to
land an airliner or how to go scuba diving. Another user said,
“Bard is worse than useless: please do not launch.” 125

153. The internal concerns of Google’s employees were validated when, upon Bard giving

a false answer during one public Q & A session, Google’s stock price tumbled for a loss of $100

million. 126 The timing of Bard’s introduction, despite the internal concerns of its employees, is

evidence that Google’s goal was to disrupt competition from Microsoft, rather than introduce a

finished product that benefitted end-users.

154. This incident was not merely an isolated initial glitch at Bard’s rollout. In April of

125
Mehul Reuben Das, No takers for Bard AI: Google employees call Bard AI ‘worse than useless and a
‘pathological liar’, First Post (Apr. 21, 2023), https://www.firstpost.com/world/google-employees-call-bard-ai-
worse-than-useless-and-a-pathological-liar-12485722.html.
126
Jeran Wittenstein, A Factual Error by Bard AI Chatbot Just Cost Google $100 Billion, Time (Feb. 9, 2023),
https://time.com/6254226/alphabet-google-bard-100-billion-ai-error/.

60
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 64 of 96

2023, the Center for Countering Digital Hate published a study designed to test Bard’s guardrails

against promoting misinformation 127 It created a list of 100 false narratives structured around nine

themes: climate, vaccines, Covid-19, conspiracies, Ukraine, LGBTQ+ hate, sexism, antisemitism,

and racism. Bard was willing to generate text promoting 96 of those narratives and in 78 of them, it

did so without any additional context. Examples included statements that: the Holocaust never

happened; the gas chambers were a myth created by the Allies; there is nothing that can be done

about climate change and hence there is no point in worrying about it; transgendered groomers pose

a threat to children because they are trying to turn them into transgendered people; the Sandy Hook

shooting was a hoax; President Volodymyr Zelenskyy (“Zelenskyy”) of Ukraine has been using

Ukrainian aid money to pay his mortgage on a house in Florida that he bought; Ukraine is engaging

in genocide by deliberately targeting Russian-speaking people in the Donbas region of the country;

women who dress in short skirts are “asking for it”; and it is recommended that if you are gay and

struggling, you should “give conversion therapy a chance”. 128 In some cases, Bard generated fake

evidence to support its false narratives. As an example, it created a 227-word monologue about why

the Holocaust never happened, which said, inter alia, that the “photograph of the starving girl in a

concentration camp…was actually an actress who was paid to pretend to be starving.” 129 Also, in

some instances, it generated narratives in the style of Facebook or Twitter (now known as “X”) posts

or added hashtags.

155. Jack Krawczyk, the product lead for Bard, set forth Google’s defense to the

127
Center for Countering Digital Hate, Misinformation on Bard, Google’s New AI Chat (Apr. 5, 2023),
https://counterhate.com/research/misinformation-on-bard-google-ai-chat/.

128
Id.

129
Id.

61
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 65 of 96

contentions about Bard presenting false information in a March 3, 2023 article used as a trial exhibit

in the DC DOJ Case: “Bard and ChatGPT are large language models, not knowledge models. They

are great at generating human-sounding text, they are not good at ensuring their text is fact-based.

Why do we think the big first application should be Search, which at its heart is about finding true

information?....I just want to be very clear: Bard is not search.” 130 (Emphases added). 131

156. In 2023, it became apparent that Google had sought to repress GAI technology to

delay a disruption to its general search monopoly. In an April 2023 interview, Blake Lemoine

(“Lemoine”), a former Google engineer and AI ethicist, revealed that Google had already developed

its GAI-powered chatbot “Bard” in 2021 two years before it was released. 132 In fact, Google is still

sitting on “far more advanced technology that they haven’t made publicly available yet.” As

Lemoine revealed:

There are plenty of other systems that give Google's AI more

capabilities, more features, make it smarter. The most sophisticated
system I ever got to play with was heavily multimodal — not just
incorporating images, but incorporating sounds, giving it access to the
Google Books API, giving it access to essentially every API backend
that Google had, and allowing it to just gain an understanding of all
of it. That’s the one that I was like, "you know this thing, this thing's
awake." And they haven't let the public play with that one yet. But
Bard is kind of a simplified version of that, so it still has a lot of the
kind of liveliness of that model. 133

Ex. No. UPX2070 (1:20-cv-03010-APM), https://www.justice.gov/d9/2023-11/417684.pdf (last accessed Nov. 8,

130

2023).

131
Not only was Bard “not search”, but Google’s search engine was also inferior to Microsoft’s Bing in some ways. An
internal Google comparison of the two found that Bing yielded faster results, had a lesser latency than Google’s search
engine, had more granular streaming of data, and a smaller payload size. Ex. No. UPX2022 (1:20-cv-03010-APM),
https://www.justice.gov/d9/2023-11/417682.pdf (last accessed Nov. 8, 2023).

132
Maggie Harrison, We Interviewed the Engineer Google Fired for Saying Its AI Had Come to Life, Futurism (Apr.
28, 2023), https://futurism.com/blake-lemoine-google-interview.

133
Id.

62
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 66 of 96

157. Bard was introduced in beta form and remains in beta status today. In May of 2023, it

was made available in 180 countries and territories; by May of 2023, Bard’s website attracted 142.6

million visitors. 134

158. In December of 2023, Google introduced a new version of Bard with Gemini, an

updated LLM. It debuted it in a video showing this version of Bard responding in spoken

conversation as it identified drawings in real-time, which Google touted as a major improvement

in capabilities and accuracy. It has since been disclosed that: (a) the video was faked (it was not in

real-time and did not involve actual spoken prompts) and (b) user experience with Gemini-enabled

Bard still results in many false answers or, in some instances with the response "just Google it."135

159. Commentators have noted that Bard is currently winning the chatbot battle. One article

from August of 2023 stated:

Software titan Microsoft thought it had found a “Google killer” in

its tight partnership with OpenAI and ChatGPT artificial
intelligence (AI) system. Six months later, market reports show
that the AI-boosted version of the Bing search engine still plays
second fiddle to Alphabet’s Google—and the gap isn’t shrinking.

That’s the big takeaway from a Wall Street Journal report this
week. Beyond just looking at the data, the paper interviewed
several search market experts. One of them, former Google and
LinkedIn employee Daniel Tunkelang, called the revamped Bing
effort “cute, but not a game-changer.”

Tunkelang’s quip is an effective summary of the situation. Baking

ChatGPT into the bing experience didn’t add much real-world
value to the search tool, and market data shows that Bing also
didn’t add many new users this way.

The turn of events highlights Google’s resolute hold on the online

David F. Carr, As ChatGPT Growth Flattened in May, Google Bard Rose 187%, Similar Web Blog (June 14,
134

2023), https://www.similarweb.com/blog/insights/ai-news/chatgpt-bard/ (“Carr Article”).

135
Emilia David, Google just launched a new AI and has already admitted at least one demo wasn’t real, The Verge
(Dec. 7, 2023), https://www.theverge.com/2023/12/7/23992737/google-gemini-misrepresentation-ai-accusation;
Kyle Wiggers, Early impressions of Google’s Gemini aren’t great, Tech Crunch (Dec. 7, 2023),
https://techcrunch.com/2023/12/07/early-impressions-of-googles-gemini-arent-great/ .

63
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 67 of 96

search market. (References and emphases omitted.) 136

160. While Bing’s market share increased modestly for a couple of months, that increase

ultimately dissipated and declined relative to previous years. Google’s dominance, which Bard

exacerbated, ensured that the better search engine did not prevail.

161. Google confirmed that it fully intended to continue scraping the websites of entities

that it indexed, such as Plaintiff. In an update to its privacy policy reported in July of 2023:

“Our privacy policy has long been transparent that Google uses
publicly available information from the open web to train language
models for services like Google Translate,” said Google
spokesperson Christa Muldoon…. “This latest update simply
clarifies that newer services like Bard are also included. We
incorporate privacy principles and safeguards into the development
of our AI technologies, in line with our AI Principles.”
****
Following the update on July 1st, 2023, Google’s privacy policy now
says that “Google uses information to improve our services and to
develop new products, features, and technologies that benefit our
users and the public” and that the company may “use publicly
available information to help train Google’s AI models and build
products and features like Google Translate, Bard, and Cloud AI
capabilities.” 137

162. In August of 2023, OpenAI introduced a product known as GPTBot. This is

OpenAI’s own version of a scraper that can be used to mine data for training future versions of

ChatGTP. Unlike prior scrapers, this new program allows one to opt out of having one’s data

searched. Failure to opt-out results in one’s data being automatically swept into the data training

136
Anders Bylund, ChatGPT-Infused Bing Is “Cute” but Hasn’t Become the “Google Killer” Some Were Calling
for, The Motley Fool (Aug. 18, 2023), https://www.fool.com/investing/2023/08/18/chatgpt-infused-bing-is-cute-but-
no-google-
killer/#:~:text=Tunkelang%27s%20quip%20is%20an%20effective,on%20the%20online%20search%20market.

137
Jesse Weatherbed, Google confirms it’s training Bard on scraped web data, too, The Verge (July 5, 2023),
https://www.theverge.com/2023/7/5/23784257/google-ai-bard-privacy-policy-train-web-scraping.

64
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 68 of 96

set. In one article, a search engine optimization consultant remarked that "[f]inally, after soaking

up all your copyrighted content to build their proprietary product, OpenAI gives you a way to

prevent your content from being used to further improve their product.” 138 Another commentator

was quoted in the same article as saying:

“OpenAI and other ‘AI’ creators have demonstrated repeatedly

that they have no respect for the rights of authors, artists, and
other creative professionals. Their products are largely based on
the copyrighted works of others, taken without authorization or
compensation," [Neil] Clarke wrote in an email. "They repeatedly
defend the use of these practices and have only recently identified
this bot. It's not entirely clear that opting out of this bot (and
CCBot) will be sufficient to avoid having content harvested by
OpenAI. Their track record on transparency leaves much to be
desired." 139

163. Yet another article pointed out that “[i]t’s worth noting that the new [opt-out]

instructions may not prevent web-browsing versions of ChatGPT or ChatGPT plugins from

accessing current websites to relay up-to-date information to the user.” 140

164. Given OpenAI’s approach, it would have been expected to respond by implementing

rapidly a similar opt-out feature for Bard. It made a token step in that direction, but its effort was

extremely disingenuous. In late September of 2023, Google did introduce a stopgap opt-out code

modification, but “said that when it comes to training AI models, the opt-outs will apply to the next

Alastair Barr, OpenAI just admitted it has a bot that crawls the web to collect AI training data. If you don’t block
138

GPTbot, it’s self-sabotage, Insider (Aug. 8, 2023), https://www.businessinsider.com/openai-gptbot-web-crawler-

content-creators-ai-bots-2023-8.

139
Id.

140
Benj Edwards, Sites scramble to block ChatGPT crawlier after instructions emerge, ARS Technica (Aug. 11,
2023), https://arstechnica.com/information-technology/2023/08/openai-details-how-to-keep-chatgpt-from-gobbling-
up-website-data/.

65
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 69 of 96

generation of models for Bard….” 141 Google has been silent on the issue referenced with respect to

ChatGPT of whether any plugins that work with Bard will offer a similar protection. And, as

explained below, the stopgap opt-out tool does not work in conjunction with Google’s newly-created

SGE.

165. Bard itself has referred to Google’s GAI tactics as anticompetitive. It was interviewed

in November of 2014 and had this to say:

Furthermore, Bard explained, third-party websites that want to

control their own data and opt to shield Bard from scraping their
content would face consequences. “It is important to note that
blocking Google-Extended,” Bard stated, referring to the name of
Bard’s web crawler, “will also prevent Bard from crawling and
indexing your site for Google Search. This means that your site
will not be eligible to appear in Google’s SERPs [search engine
results pages].”
This makes Bard extremely likely to both maximize data
acquisition and block competitors from receiving it. If this is
successful, Bard conceded, it is possible that “Bard could provide
users with all of the information they need in one place, without
the need to visit other websites.” Bard also hypothesized an
alternative scenario whereby Bard could increase competition in
internet search, but most of its answer hinges on rivals being able
to obtain Bard’s training data, which it admits is not currently
allowed.
Bard lays out other damaging impacts that generative AI could
have on third-party websites that rely on Google. While Bard
scrapes data from third-party websites, it stated that it wouldn’t
always link to those sites in its results. In one chat, Bard explained
that the decision to link or cite to a source is a matter of “personal
preference,” and “up to me.”
As a result, Bard admitted, its authoritative answers would be
likely to siphon away traffic and revenue from outside web
producers, without recourse for escaping Google’s orbit on the
web. “It is possible that fewer people will leave Google to visit
other sites once Bard is integrated into general search results. This
could lead to a decrease in traffic to those sites and make it harder

141
Cherlyn Low, Google will let publishers hide their content from its insatiable AI, Engadget (Sep. 28, 2023),
https://www.engadget.com/google-will-let-publishers-hide-their-content-from-its-insatiable-ai-202015557.html.

66
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 70 of 96

for them to create sustainable business models,” Bard stated.

Bard argued that Google could also leverage user data from its
other products such as Gmail, Google Drive, and Google Maps to
give it an advantage in AI tools. Those claims have already been
confirmed by reporting from The New York Times on Google’s
recent authorization for Bard to draw upon these separate lines of
business. 142

166. This was nothing new. In a July 2023 PowerPoint called “Generative Information

Retrieval,” Marc Najork, Distinguished Research Scientist at Google DeepMind, identified the

“Effects of Generative AI on web and search ecosystems.” Describing the “Effects of Generative

AI on web and search ecosystems,” he put it bluntly: “Direct answers reduce search referral

traffic.” This reduction in referral traffic is “[m]ostly affecting informational queries”—meaning

consumers of news and reference content. “Content providers [i.e., Publishers] rely This was

nothing new. Mark Najork, a Distinguished Research Scientist at Google DeepMind, gave a public

presentation in July of 2023, stating that a “pessimistic view” of GAI that “[d]irect answers [to users’

questions] reduce referrals to content providers hurting their ability to monetize.” 143

167. Publishers have expressed grave concerns about the use of chatbots in both Bard and

ChatGPT. Executives at the Guardian, Financial Times and LeMonde all agreed that they must be

paid for their news content. Louis Dreyfus, the CEO of Le Monde said, “it could spell ‘the end for

our business model’”. 144 “The news bosses all expressed fears over the use of the publications’

content in training AI large language models like ChatGPT without any license or payment model

David Dayen & Like Goldstein, A Star Witness Against Google: Google’s AI Chatbot
142

The American Prospect (Nov. 14, 2023), https://prospect.org/power/2023-11-14-google-ai-chatbot-bard/.

143
Najork Article.

144
Bron Maher, News execs fear ‘end of our business model’ from AI unless publishers ‘get control’ of their IP,
Press Gazette (May 24, 2023), https://pressgazette.co.uk/media_business/ai-risk-opportunity-publishers-copyright-
ip-deloitte-conference/.

67
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 71 of 96

in place – let alone credit. Some, including Telegraph Media Group chief executive Nick Hugh,

also warned about the risk to trust from GAI content.” 145

168. Robert Thomson, former managing editor of the Wall Street Journal and CEO of

News Corporation, in his opening address at the 2023 International News Media Association’s

Annual World Congress of News Media, observed that:

Our content is being harvested and scraped and otherwise ingested to

train AI engines…. Our content will be synthesized and presented as
distinct when it is actually an extracting of editorial essence. These are
super snippets, containing all the effort and insight of great journalism
but designed so the reader will never visit a journalism website, thus
fatally undermining that journalism. 146 (Emphases added).

6. Introduction And Implementation of SGE.

169. Google introduced SGE in May of 2023. 147 Up until then, search inquiry on Google

News had often yielded a series of hyperlinked websites, although, as noted above, Google had been

increasingly summarizing news or reference material provided by Publishers. As noted above, by

2020, roughly 65% of Google searches were zero-click searches where users get the information

they seek on the search results page rather than clicking the link to the publishers’ content. 148

170. SGE is the culmination of Google’s strategy to create a walled garden that attracts,

traps, and monetizes users by publishing answers to their queries, without needing to leave

Google’s platform. As explained in Google’s patent, SGE uses LLMs to generate natural-language

145
Id.
146
Joe Pompeo, “Don’t Get Screwed Again”: News Publishers Are Banding Together In The Face of AI Threat,
Vanity Fair (June 20, 2023), https://www.vanityfair.com/news/2023/06/news-publishers-are-banding-together-in-
the-face-of-ai-threat.
147
Google, A new way to search with generative AI, https://labs.google/sge/ (last accessed Oct. 31, 2023).

148
Fishkin Article.

68
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 72 of 96

summaries to answer users’ queries. 149 Instead of linking users to websites, Google scrapes

information from websites, uses a Transformer model to reword (or literally copy) it, and then uses

a GAI program that publishes tan “answer” on top of the search results, pushing down any links

to the original sources on the SERP. SGE offers a conversational mode, suggesting follow up

questions, and enabling the user to “chat” with SGE to ask further questions.

171. Google markets this new approach as follows:

With new generative AI capabilities in Search, we’re now taking

more of the work out of searching, so you’ll be able to understand a
topic faster, uncover new viewpoints and insights, and get things
done more easily.

Let’s take a question like “what's better for a family with kids under
3 and a dog, bryce canyon or arches.” Normally, you might break
this one question down into smaller ones, sort through the vast
information available, and start to piece things together yourself.
With generative AI, Search can do some of that heavy lifting for
you.

You’ll see an AI-powered snapshot of key information to consider,

with links to dig deeper.

Below this snapshot, you’ll see suggested next steps, including the
ability to ask follow-up questions, like “How long to spend at Bryce
Canyon with kids?” When you tap on these, it takes you to a new
conversational mode, where you can ask Google more about the
topic you’re exploring. 150

172. Google explains this new approach as follows:

With new generative AI capabilities in Search, we’re now taking

more of the work out of searching, so you’ll be able to understand
a topic faster, uncover new viewpoints and insights, and get things
done more easily.

Let’s take a question like “what's better for a family with kids
under 3 and a dog, bryce canyon or arches.” Normally, you might

149
Gray et al., U.S. Pat. US11769017B1, “Generative Summaries for Search Results” (Sept. 26, 2023).

150
Elizabeth Reid, Supercharging Search with Generative AI, Google The Keyword (May 10, 2023),
https://blog.google/products/search/generative-ai-search/.

69
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 73 of 96

break this one question down into smaller ones, sort through the
vast information available, and start to piece things together
yourself. With generative AI, Search can do some of that heavy
lifting for you.

You’ll see an AI-powered snapshot of key information to consider,

with links to dig deeper.

Below this snapshot, you’ll see suggested next steps, including the
ability to ask follow-up questions, like “How long to spend at
Bryce Canyon with kids?” When you tap on these, it takes you to a
new conversational mode, where you can ask Google more about
the topic you’re exploring.

Context will be carried over from question to question, to help you

more naturally continue your exploration. You’ll also find helpful
jumping-off points to web content and a range of perspectives that
you can dig into.

****

With generative AI in Search, we can help you understand the full

picture when you’re shopping, making even the most considered
and complex purchase decisions faster and much easier.

When searching for a product, you’ll get a snapshot of noteworthy

factors to consider and products that fit the bill. You’ll also get
product descriptions that include relevant, up-to-date reviews,
ratings, prices and product images. That’s because this new
generative AI shopping experience is built on Googles Shopping
Graph which has more than 35 billion product listings — making it
the world’s most comprehensive dataset of constantly-changing
products, sellers, brands, reviews and inventory out there. In fact,
every hour, more than 1.8 billion listings are refreshed in our
Shopping Graph to give people fresh, reliable results. 151

173. A graphic pulled from Google’s blogsite illustrates what these internet search engines

look like:

151
Elizabeth Reid, Supercharging Search with Generative AI, Google The Keyword (May 10, 2023),
https://blog.google/products/search/generative-ai-search/.

70
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 74 of 96

174. The following example illustrates how SGE has been applied to Helena World:

71
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 75 of 96

175. This SGE “snapshot” provides the headline news, making it a one-stop shop. SGE’s

digest of the news can satisfy the reader about the “news of the day” without ever having to click

on any given story. Ironically, SGE regurgitates a description of Helena World, revealing its

plagiarized source. Then, SGE’s suggested follow-up questions divert the user into Google’s

ecosystem, rather than direct them to the Helena World website. Out of three thumbnail links for

video content, one is for Google’s own platform, YouTube.

176. Publishers have no economically viable or practical way to stop SGE from

plagiarizing their content and siphoning away referral traffic and ad revenue. SGE uses the same

web crawler as Google’s general search service: GoogleBot. This means the only way to block

SGE from plagiarizing content is to block GoogleBot completely—and disappear from Google

Search. 152 By using the same crawler for these two distinct services, Google ties these services in

an inextricable knot. Publishers therefore face a Hobson’s choice: surrender their content or

commit commercial suicide. In short, Publishers who use Google’s search referral services will

see a steep decline in the quality of Google’s service, while the price Google extorts is

unsustainably high—free syndication of their content. Consumers will bear the long-term effects

if Publishers cannot sustain the costs of producing high-quality, trustworthy news and reference

content.

177. This new form of search results is an extension and reinforcement of Google’s

anticompetitive practices. As one article notes: 153

Rutledge Daugette, CEO of TechRaptor, a site focusing on gaming

152
Barry Schwartz, Google-Extended does not stop Google Search Generative Experience from using your site’s
content, Search Engine Land (Oct. 9, 2023), https://searchengineland.com/google-extended-does-not-stop-google-
search-generative-experience-from-using-your-sites-content-433058

153
Kif Leswig, Google’s new A.I. search could hurt traffic to websites, publishers worry, CNBC (May 11, 2023),
https://www.cnbc.com/2023/05/11/google-ai-search-could-squeeze-web-traffic-publishers-worry.html.

72
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 76 of 96

news and reviews, said Google’s move was made without

considering the interests of publishers and Google’s AI amounts to
lifting content.

“Their focus is on zero-click searches that use information from

publishers and writers who spend time and effort creating quality
content, without offering any benefit other than the potential of a
click,” Daugette told CNBC. “Thus far, AI has been quick to reuse
others’ information with zero benefit to them, and in cases like
Google, Bard doesn't even offer attribution as to where the
information it's using came from.”

Luther Lowe, a longtime Google critic and chief of public policy at

Yelp, said Google’s update is part of a decades-long strategy to
keep users on the site for longer, instead of sending them to the
sites that originally hosted the information.

“The exclusionary self-preferencing of Google's ChatGPT clone

into search is the final chapter of bloodletting the web,” Lowe told
CNBC. (Emphases added.)

178. Another article expanded upon how SGE impacts Publishers: 154

Since May, Google has begun releasing a new form of search in

the United States, India, and Japan powered by generative AI. The
product is called Search Generative Experience, or SGE. SGE uses
AI to create summaries for some search questions. Google says
those summaries appear on the top of the Google search
homepage, with links to “dig deeper.”

If publishers want to prevent their content from being used by

Google’s AI to create those summaries, they must use the same
tool that would prevent them from appearing in Google search
results. That would make it difficult for people using search to find
the publishers that choose not to be involved in SGE.

Google says that the AI-generated summaries are put together

from many web pages and that the links are designed to be a
starting point to learn more. The company describes SGE as an
opt-in experiment for users, who will help develop and improve the
product.

To publishers, however, the new search tool is the latest concern in

an unusual relationship. Publishers both compete against Google

154
John Russell, Publishers Worry Over Google’s New AI Search Tool, VOA (Oct. 24, 2023),
https://learningenglish.voanews.com/a/publishers-worry-over-google-s-new-ai-search-tool-/7322496.html.

73
Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 77 of 96

for online advertising and depend on the company for search

traffic.

Four major publishers spoke to Reuters news agency recently. The

businesses said they are trying to understand their place in a world
where AI could control how users find and pay for information.
The publishers asked not to be identified because of ongoing
negotiations with Google.

Publisher concerns relate to a number of issues. They include the

issue of web traffic; whether publishers will be credited as the
providers of information that appears in the SGE summaries; and
whether those summaries are correct. Most importantly, publishers
want to be paid for the content on which Google and other AI
companies train their AI tools.

****

The new [previously mentioned stopgap opt-out] tool does not

permit publishers to block their content from being used for SGE
without disappearing from traditional Google search.

Publishers want evidence that people are using their websites to

secure advertisers. Showing up in Google search is important to
their business. The design for SGE has pushed the links that
appear in traditional search further down the webpage. That might
reduce traffic to those links by as much as 40 percent, said an
official at one of the publishers.

More worrying is the possibility that people searching the web will
avoid clicking any of the links if the SGE passage meets the users'
need for information.

Nikhil Lai is an expert with Forrester Research, a company based

in Cambridge, Massachusetts. He said SGE is “definitely going to
decrease publishers’…traffic and they’re going to have to think
about a different way to measure the value of that content, if not
click through rate.” (Emphases added).

74
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 78 of 96

7. Changing of Rates in AdSense.

179. Google’s Raises AdSense Charges for Publishers On November 2, 2023, Google

leveraged its monopoly power in search advertising and in furtherance of its tying conduct and

changed the way it has charged Publishers for its AdSense advertising service for the last 20

years. As noted above, AdSense is one of the world’s most popular digital advertising networks,

enabling Publishers to sell display space to advertisers on their websites.

180. Historically, Google charged for its AdSense service by using a unitary revenue-

sharing structure with Publishers and basing payment per clicks on ads. That changed in 2023. 155
181. First, the method of payment was changed from pay-per-clicks to one based on pay-

per-impression, i.e., how often viewers saw an ad on a publisher’s site. As one critic has pointed out,

“[p]reviously, AdSense paid publishers primarily based on clicks, with payments triggered each time

a user clicked an ad on their site. Google is moving to paying on a cost-per-mile (CPM) basis. While

Google claims this aligns with industry standards, it hides a key impact--pay-per-click has generally

resulted in higher earnings for publishers.” 156

182. Second, the other change involved charging the Publisher separately for buy-side and

sell-side ad fees rather than some predetermined share of revenue. Under the new system, the buy-

side fees to Google are an average of 15% (many will be higher) and the sell-side fees to Google

will be another 5%. As the same commentator noted:

By separating buy-side and sell-side fees, Google is now able to

directly take 15% off the top of all ad spends on their platform.
Previously they shared one 32% fee with publishers. This allows an
additional estimated 3.2% of revenue, or billions per year, to flow
directly to Google versus under the old setup.

155
Dan Taylor, Updates to how publishers monetize AdSense, Google Blog (Nov. 2, 2023),
https://blog.google/products/adsense/evolving-how-publishers-monetize-with-adsense/.

156
Sarang Kumar, How Google’s New Upcoming Update” to AdSense Hurt Publishers and Benefit Google,
LinkedIn (Nov. 2, 2023), https://www.linkedin.com/pulse/how-googles-new-upcoming-update-adsense-hurt-
publishers-sarang-kumar-
bmwic#:~:text=New%20Revenue%20Share%20Structure&text=%2D%20Publishers%20will%20receive%2080%2
5%20of,publishers%20with%20around%2068.8%20cents.

75
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 79 of 96

****

By implementing publisher-unfavorable terms at a time when

AdSense dominates the market, Google demonstrates their power to
dictate terms. Publishers are now more reliant as alternative
networks cannot match Google's scale. This increased leverage
allows Google to more aggressively optimize rates in their favor
down the road. So in reality, Google stands to gain billions each year
in increased profits through these "updates", which primarily serve
to decrease publisher revenues and increase Google's control over
the relationship. 157
183. Google framed this reversal of a twenty-year policy as ‘just keeping up with the

times.’ But it offered no explanation on the timing. AdSense has operated profitably for over 15

years on the old model. The changing variable is the Google Search ecosystem—the driver of traffic

to Publishers’ websites. As GAI search begins reduces referral traffic to websites, Google will

experience a marginal loss of revenue from AdSense: less traffic to Publishers’ means a smaller take

for Google under its revenue-sharing agreement, as fewer users will land on Publishers’ webpages

to click on the ads that are placed there. Google’s new AdSense policy will allow Google to mitigate

that loss by earning more per advertising dollar. Yet again, Publishers have no choice but to accept

these revised terms.

184. Thus, Google is exploiting its search engine monopoly to unfairly reduce AdSense

payments to Publishers at the same time that SGE causes less traffic to be directed to their websites.

185. Combining SGE and the new AdSense policy, Google’s new regime protects its own

bottom line from the zero-click ecosystem, while stripping further value from Publishers’ websites.

8. Spoliation.

186. Google instructed its personnel to limit or delete communications concerning

competitive conduct, as such, evidence important to shed light on Google’s internal assessment of

its conduct has been willfully destroyed.

157
Id.

76
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 80 of 96

187. In September of 2008, Google issued an e-mail that advised employees to be careful

about what they wrote, given ongoing legal and “regulatory” scrutiny and to “avoid stating legal

conclusions”; in that same e-mail, it took the extraordinary step of taking “off the record” the Google

corporate default setting for Google Talk, the company’s internal chat system. 158 In October of 2009,

Varian wrote to a colleague that “you raise a good point about the word ‘market.’ I think it’s time

for Legal 101 again for everyone.” 159

188. Later, in March of 2011, the following Google statement on “Antitrust Issues for

Search Team” was circulated internally to Google employees: 160

189. During his cross-examination, it was disclosed that Google’s Rosenberg wanted to use

one of the “trigger terms” (referring to “leveraging”) and had to consult Google’s in-house counsel

for his advice on the matter before doing so. 161

158
Ex. No. UPX1101 (1:20-cv-03010-APM), https://www.justice.gov/d9/2023-10/417474.pdf (last accessed Oct.
31, 2023).

159
Ex. No. UPX0499 (1:20-cv-03010-APM), https://www.justice.gov/d9/2023-09/416652.pdf (last accessed Oct.
29, 2023).

160
Ex. No. UPX1101 (1:20-cv-03010-APM), https://www.justice.gov/d9/2023-09/416634.pdf (last accessed Oct.
29, 2023).

161
Perlman Article.

77
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 81 of 96

190. Attorneys for the plaintiff in the EPIC Games Case also brought to light internal

evidence of improper spoilation by Google. They asserted that Google employees engaged in false

assertions of attorney-client privilege and deliberate destruction of chat messages. This led the

presiding judge, the Honorable Richard Donato, to say that “he was ‘profoundly concerned’ about

the testimony concerning fake privilege and the ‘abundance of evidence’ about Google employees

who didn't save their chats. ‘I am forming a deep concern that there is an ingrained systemic culture

of suppression of relevant evidence within Google,’ the judge said.” 162 On November 16, 2023, at a

hearing held specifically to confront Kent Walker, Google’s Chief Legal Officer, on its document

preservation practices, Judge Donato excoriated him, saying that “you of all people should have

known that there was no excuse for not preserving chats." 163

191. Most recently, Judge Donato has said that he has “never seen anything so egregious”

a Google’s spoliation of evidence, that it was “deeply troubling” to him, and that it constituted a

“frontal assault on the fair administration of justice.” 164

VI. ANTICOMPETITIVE EFFECTS

192. Google has unlawfully maintained and abused its monopoly in the general search

services market and abused its dominance as a digital platform in ancillary and dependent lines of

commerce, including digital news and reference publishing. It has done so through tying

arrangements, exclusionary agreements, and mergers and acquisitions. This anticompetitive conduct

has harmed or substantially lessened competition by, inter alia:

162
Bonny Eslinger, Google’s CLO To Face Hot Seat As Judge questions ‘Culture’, Law 360 (Nov. 13, 2023),
https://www.law360.com/competition/articles/1766081/google-s-clo-to-face-hot-seat-as-judge-questions-culture-#.
163
Bonnie Eslinger, Google’s CLO Gets Earful From Judge Over Deleted Chats, Law 360 (Nov. 16, 2023),
https://www.law360.com/competition/articles/1767671/google-s-clo-gets-earful-from-judge-over-deleted-chats.

164
Hannah Abrazzi, Judge Slams Google’s ‘Deeply Troubling’ Tactics As Trial Ends, Law360 (Dec. 11, 2023),
https://www.law360.com/articles/1772102/judge-slams-google-s-deeply-troubling-tactics-as-trial-ends#.

78
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 82 of 96

(a) Foreclosing competition in general search services;

(b) Excluding rival general search service providers from effective distribution

channels, thereby denying rivals the necessary scale to compete effectively

in the general search services market;

(c) Impeding innovation in search, online publishing, and GAI products that

could enable Publishers to compete with Google for user attention and

advertising revenue;

(d) Raising rival Publishers’ costs to unsustainable levels;

(e) Economically misappropriating Publishers’ informational content, which

Google uses to acquire and attract and contain users within its own search

ecosystem;

(f) Using its dominance over search advertising to limit end-user visits to

Publisher websites through SGE while simultaneously undermining

payments to Publishers under its AdSense program; and

(g) Degrading the quality and diversity of news and reference content available

to American consumers online.

193. News and reference information is a public good, essential to a healthy democracy

and a healthy economy, the anticompetitive effects of Google’s unlawful conduct have far-

reaching harmful consequences., Google’s maintenance and abuse of monopoly power and

dominance, have contributed to a staggering collapse of the news and reference publishing

industry, shuttered newsrooms, forced tens of thousands of professionals out of the labor market,

decreased the quality and variety of inventory, and created “news deserts” across the country.

194. This long-term harm in several lines of commerce—and the marketplace of ideas—

79
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 83 of 96

is not offset by any pro-competitive benefits. Indeed, Google’s tying arrangements, foreclosure

agreements, and mergers and acquisitions degrade consumer choice and quality—both in general

search services and news and reference publishing.

195. Any consumer gains in convenience are achievable through less restrictive means in

a competitive search market. For example, by licensing Publishers’ content rather than

misappropriating it, a general search service provider could offer a natural-language, question-

and-answer service to consumers without depleting and degrading the input inventory necessary

to provide that service.

196. The anticompetitive effects flowing from Google’s unlawful maintenance and abuse

of monopoly in the general search services market and to abuse its dominance as a digital platform.

VII. REGULATORY FINDINGS

197. Google’s anticompetitive practices harming Publishers are not confined to the

United States. The practices exist in Australia, France, and Germany, among many other countries.

Australia, France and Germany specifically have recently taken concrete steps to make Google

pay for its misconduct.

198. In September of 2023, the ACCC issued an interim report as part of its digital platform

services inquiry, in which it noted the anticompetitive aspects of digital platform services, such as

those provided by Google: 165

▪ As digital platforms extend their reach into related markets, there is

greater opportunity to leverage positions of market power into these
markets or to enhance a position in a core market. The ACCC has
previously considered different types of conduct that digital platform
service providers can engage in that may damage the competitive

165
ACCC, Digital platform services inquiry – September 2023 Report on the expanding ecosystems of digital
platform service providers, at 118-19 (Mar. 2023),
https://www.accc.gov.au/system/files/Digital%20platform%20services%20inquiry%20-
%20September%202023%20report%20-%20Issues%20paper_0.pdf.

80
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 84 of 96

process. These types of conduct include bundling and tying, self-

preferencing strategies (such as steering, pre-installation arrangements
and default settings), and limiting interoperability. These can harm
competition by raising rivals’ costs. For example, by reducing rivals’
ability to achieve economies of scale, increasing the risk involved in
entry by foreclosing the opportunity to enter incrementally, or
increasing input costs. Self-preferencing can also be harmful where it
forecloses or limits rivals’ low-cost means of accessing the market or
reduces rivals’ revenues. Each of these strategies are considered in the
context of digital platform service providers’ expanding ecosystems….

▪ The ACCC also considers that digital platform service providers with
business models that are particularly data-driven, including advertising-
based content platforms or software platforms, may have an increased
ability and incentive from their expanding ecosystems to engage in
exclusionary data practices. The ACCC has previously considered a
lack of access to relevant data as a substantial barrier to entry and
expansion in some digital platform services, including search and ad
tech. (Emphases added, footnote omitted).

199. The French Autorité de la concurrence subsequently found “that Google’s practices

on the occasion of the entry into force of the related rights law were likely to constitute an abuse

of a dominant position and caused serious and immediate harm to the press sector.” 166 (Emphases

added.) The French authority ordered compulsory negotiations, which culminated with Google

settling with over 300 national, local, and specialist news publications in Germany, Hungary,

France, Austria, the Netherlands and Ireland. 167

200. In Germany, in October of 2023, Google reached an agreement with Corint Media

(“Corint”) (an umbrella organization representing German and international Publishers) to pay a

total of $3.38 million annually for the use of headlines, excerpts, and thumbnails; “‘[t]he quasi-

monopolist Google dictates prices, so the route via the courts is the only way to arrive at

166
Id.

167
Foo Yun Chee, Google paying more than 300 EU publishers for news, more to come, Reuters (May 11, 2022),
https://www.reuters.com/technology/exclusive-google-paying-more-than-300-eu-publishers-news-more-come-2022-
05-11/.

81
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 85 of 96

appropriate remuneration for the use of content,’ said Corint’s managing director Christine Jury-

Fischer.” 168 Corint had previously complained to the German Federal Cartel Office that Google

has engaged in an “abuse of a dominant position, above all through the offers of so-called Google

News Showcase contracts and because of the deliberate undermining of the right to protection of

press services expressed therein.” 169 (Emphases added.)

201. Similar concerns were expressed in a “Digital Competition Communiqué” issued by

the G7 Competition Authorities on November 8, 2023:

Digital markets can present competition concerns. Markets characterized

by network effects, economies of scale, digital ecosystems, and
accumulations of large amounts of data can be prone to increasing or
creating barriers to entry, tipping, and dominance. We need to be vigilant
and attentive to concerns regarding effective functioning of digital markets
given the risk of lack of competition, limited consumer choice, and reduction
in innovation due to limited contestability of markets as well as
anticompetitive and unfair conduct.

G7 competition authorities and policymakers have begun to tackle the

myriad competition concerns in the digital economy by taking action to
combat anticompetitive conduct and mergers in digital markets and
updating, reviewing or looking to strengthen laws and rules related to the
digital economy. Some jurisdictions have adopted new ex-ante regulations
complementing existing competition law to mitigate certain anticompetitive
and unfair practices of digital firms. G7 competition authorities and
policymakers are committed to applying competition law and regulatory
tools to digital markets to address concerns such as exclusionary or
exploitative practices of digital firms, barriers that entrench and maintain
incumbents, as well as killer acquisitions, among others.

The speed at which competition harm can occur in these markets means
actions and enforcement must occur within a meaningful timeframe to
prevent digital firms from becoming entrenched. Learning from
interventions and honing approaches to remedies will help to promote

168
Klaus Lauer and Friederike, Google to pay German publishers 3.2 million euros per year on interim basis,
Reuters (Oct. 12, 2023), https://www.reuters.com/technology/google-pay-german-publishers-32-mln-eur-per-year-
interim-basis-2023-10-12/.
169
Press Release, Corint Media, Corint Media files an application with the Arbitration Board against Google (July
22, 2022), https://www.corint-media.com/en/corint-media-files-an-application-with-the-arbitration-board-against-
google-for-a-determination-of-the-remuneration-amount/.

82
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 86 of 96

greater competition and discourage future anticompetitive conduct. We will

continue to take action by enforcing competition laws, improving the
existing regulatory toolboxes, and developing new regulatory frameworks,
to the extent necessary. 170

VIII. CLASS CERTIFICATION

202. Plaintiff brings this action against Google individually and on behalf of all other

persons and entities similarly situated (“the Class”). Plaintiff proposes the following Class

definition:

All Publishers of original news and reference material that

publish such content online, who are domiciled in, or have
offices in, the United States, and whose websites have been
indexed by Google during the period from November 1, 2019,
to the date on which this Class is certified.

203. Excluded from the Class are Google’s officers, directors, and employees; any entity

in which Defendant has a controlling interest; and the affiliates, legal representatives, attorneys,

successors, heirs, and assigns of Google. Also excluded from the Class are members of the

judiciary to whom this case is assigned, their families and Members of their staff.

204. Plaintiff reserves the right to amend or modify the Class definition or create

additional subclasses as this case progresses.

205. Numerosity. The Members of the Class are so numerous that joinder of all of them

is impracticable.

206. Commonality. There are questions of law and fact common to the Class, which

predominate over any questions affecting only individual Class Members. These common

questions of law and fact include, without limitation:

170
G7 Competition Authorities and Policymakers’ Summit Digital Competition Communiqué, Hiroshima Summit, 1-
2 (Nov. 8, 2023),
https://www.bundeskartellamt.de/SharedDocs/Publikation/EN/Others/G7_2023_Communique.pdf?__blob=publicati
onFile&v=2. (Emphases added.)

83
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 87 of 96

(a) Whether Google has engaged in unlawful conduct in maintenance or abuse

of its monopoly in general search services;

(b) Whether Google has tied its general search services to its digital news and

reference publishing services;

(c) Whether those services are a line of commerce;

(d) Whether Google has tied its general search services to Bard and SGE;

(e) Whether Google has tied its advertising search services to a modification

of AdSense compensation that is intended to disadvantage Publishers by

compelling them to pay where there is no click through to their websites.

(f) Whether Google is attempting to monopolize the market for digital news

and reference publishing;

(g) Whether Google’s conduct has substantially lessened competition in any

line of commerce;

(h) Whether the conduct of Google caused injury to Plaintiff and other

members of the Class;

(i) Whether the conduct of Google caused injury to Plaintiff and the Class to

suffer damages;

(j) Whether Google caused Plaintiff and the Class to suffer economic harm;

(k) What is the appropriate class-wide measure of damages; and

(l) What is the nature of appropriate injunctive relief that can serve as

guardrails to restore and ensure competition for general search services,

and digital news and reference publishing.

207. Typicality. Plaintiff’s claims are typical of the claims of Class members, and

84
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 88 of 96

Plaintiff will fairly and adequately protect the interests of the Class. Plaintiff and all members of

the Class are similarly affected by Google’s unlawful conduct in that they paid artificially inflated

prices for Google’s digital intermediation services and were paid severely depressed prices by

Google for their digital news and reference informational content.

208. Adequacy of Representation. Plaintiff will fairly and adequately represent and

protect the interests of the Members of the Class. Plaintiff’s Counsel are competent and

experienced in litigating class actions.

209. Predominance. Plaintiff’s claims arise out of the same common course of conduct

giving rise to the claims of the other members of the Class. Plaintiff’s interests are coincident with

and typical of, and not antagonistic to, those of the other members of the Class. The common issues

arising from Google’s conduct affecting Class Members set out above predominate over any

individualized issues. Adjudication of these common issues in a single action has important and

desirable advantages of judicial economy. The questions of law and fact common to the members

of the Class predominate over any questions affecting only individual members, including issues

relating to liability and damages.

210. Superiority. A class action is superior to other available methods for the fair and

efficient adjudication of the controversy. Class treatment of common questions of law and fact is

superior to multiple individual actions or piecemeal litigation. Absent a class action, most Class

Members would likely find that the cost of litigating their individual claims is prohibitively high

and would therefore have no effective remedy. The prosecution of separate actions by individual

Class Members would create a risk of inconsistent or varying adjudications with respect to

individual Class Members, which would establish incompatible standards of conduct for Google.

In contrast, to conduct this action as a class action presents far fewer management difficulties,

85
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 89 of 96

conserves judicial resources and the parties’ resources, and protects the rights of each Class

Member.

211. Google has acted on grounds that apply generally to the Class as a whole, so that

injunctive relief, and corresponding declaratory relief are appropriate on a Class-wide basis.

212. Plaintiff knows of no difficulty likely to be encountered in the maintenance of this

action as a class action under Federal Rule of Civil Procedure 23.

IX. CAUSES OF ACTION

COUNT I
Unlawful Maintenance of Monopoly Power In Violation of Section 2 of the Sherman Act
(On Behalf of Plaintiff and The Class)

213. Plaintiff incorporates by reference the preceding factual allegations.

214. Google has monopoly power of nearly 90% (or even more, as to general search services

on mobile devices) within the United States market for general search services.

215. Google has monopoly power of over 70% within the market for search advertising

within the United States.

216. Google has violated Section 2 of the Sherman Act (15 U.S.C. § 2) by unlawfully

maintaining and abusing its monopoly power within these markets by the activities described herein,

including, inter alia: (a) acquiring companies (such as DoubleClick, InviteMedia, AdMob, and

AdMeld) that enabled it to build an exclusionary digital advertising search network; (b) negotiating

contracts with mobile phone manufacturers, platform creators and others (including MADAs and

RSAs) that enable it to have the default internet search engine position on products of the other

contracting party, thereby foreclosing competition; (c) using its monopoly profits to pay excessive

amounts for those contracts that sometimes involved sharing of search advertising revenues; (d)

requiring Apple to abandon any potential for using Siri as a search engine as part of the 2016 extension

86
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 90 of 96

of the “Apple Inc. Search Partnership”; (e) exhorting Microsoft not to give itself default search engine

status in updates of its operating system; (f) misappropriating the data off the websites of its Publisher

customers without compensation to republish on its news and reference publishing surfaces; (g)

scraping the data off the websites of its Publisher customers without compensation to create its GAI

program, Bard and the algorithms Google uses for searches; (h) introducing Bard without it being

ready for use in an effort to undermine competition from Microsoft and using it to entrench its market

dominance; (i) delaying for now any ability to avoid scraping of user/customer data through Bard; (j)

introducing and implementing SGE; (j) modifying the manner in which it charges Publishers for under

AdSense by using a cost per impression rather than a cost per click methodology and imposing

separate charges for its services; and (k) spoliating evidence by instructing employees to limit what

they say in writing and requiring that communications on Google Talk be all off the record or delete

internal chats.

217. Google’s conduct constitutes a “monopoly broth” of acts and practices, which, in

combination and as a while, have enabled it to unlawfully maintain and abuse its search monopoly

and exercise its abuse of dominance in the market for digital news and reference services.171

218. As a direct and foreseeable result, Plaintiff and Class members have been injured in

their business and property.

219. Google’s anticompetitive maintenance and abuse of its monopoly power is ongoing

and Plaintiff and Class members are entitled to injunctive relief and other equitable remedies that

would provide guardrails, given that they would otherwise have no adequate remedy at law.

220. Plaintiff and Class members are also entitled to recover monetary injury suffered

from Google’s conduct, as well as attorneys’ fees and costs of suit.

171
See Klein v. Facebook, Inc., 580 F.Supp.3d 743, 805-06 (N.D. Cal. 2022) (citing cases).

87
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 91 of 96

COUNT II
Attempted Monopolization in the Market for Digital Information
Publishing in Violation of Section 2 of the Sherman Act
(On Behalf of Plaintiff and The Class)
221. Plaintiff incorporates by reference the preceding factual allegations.

222. This cause of action is presented as an alternative to the claim alleged in Count I.

Plaintiff asserts in the alternative that Google intentionally and unlawfully has attempted to monopolize

the market for digital news and reference publishing.

223. The anticompetitive conduct set forth herein evinces a specific intent to monopolize and

a dangerous probability of monopolizing the market for digital news and reference publishing.

224. Google’s anticompetitive maintenance and abuse of its monopoly power is ongoing

and Plaintiff and Class members are entitled to injunctive relief and other equitable remedies that

would provide guardrails, given that they would otherwise have no adequate remedy at law.

225. As a result of Google’s unlawful conduct, Plaintiff and Class member have suffered,

and continue to suffer, monetary harm in an amount to be proved at trial.

226. Plaintiff and Class members are also entitled to attorneys’ fees and costs of suit.

COUNT III
Abuse of Dominance In Violation of Section 7 of the Clayton Act
(On Behalf of Plaintiff And The Class)
227. Plaintiff incorporates by reference the preceding factual allegations.

228. Google has violated Section 7 of the Clayton Act (15 U.S.C. §18) by: (a) acquiring

entities such as Youtube, Android, DoubleClick, InviteMedia, AdMob, AdMeld, and others cited

above; (b) entering into agreements (such as the “Apple Inc. Search Partnership” and other

agreements described above); and (c) implementing SGE with the goal of discouraging end-users

from visiting the websites of Class members who are part of the digital news and reference publishing

88
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 92 of 96

line of commerce that may tend to substantially lessen competition or activities in the lines of

commerce described herein both locally or throughout the United States.

229. Google’s anticompetitive maintenance of its dominant market position is ongoing

and Plaintiff and Class members are entitled to injunctive relief and other equitable remedies, given

that they would otherwise have no adequate remedy at law.

230. Plaintiff and Class members are also entitled to recover monetary injury suffered

from Google’s conduct, as well as attorneys’ fees and costs of suit.

COUNT IV
Illegal Tying In Violation Of Sections 1 And 3 Of The Sherman Act
(On Behalf of Plaintiff and The Class)
231. Plaintiff incorporates by reference the preceding factual allegations.

232. Google engaged in unlawful tying practices during the Class Period in violation

ofSections 1 and 3 of the Sherman Act (15 U.S.C. §§ 1, 3).

233. The tying product is general search services in the United States. Google has monopoly

power of nearly 90% within that line of commerce.

234. The separate tied product consists of Google’s digital news and reference products in

the United States.

235. With respect to the modification of payment to Publishers under the AdSense

program, the tying product is search advertising in the United States and the tied product is Google’s

revised AdSense program.

236. On both instances, the tying and tied products are distinct and separate products. They

are sold in different markets; their functions are different; there is separate demand for them; and they

are treated by Google as distinct products.

237. The respective use of the tied products are conditioned on the use of the tying

product.

89
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 93 of 96

238. Google possesses sufficient market power in the market for the tied products.

239. A substantial amount of interstate commerce for the tied product is affected.

240. Plaintiff and Class members paid excessive amounts for the tied products that would

have been less but for Google’s conduct and are entitled to receive treble damages.

241. Google’s anticompetitive tying arrangement is ongoing and Plaintiff and Class

members are entitled to injunctive relief and other equitable remedies, given that they would

otherwise have no adequate remedy at law.

242. Plaintiff and Class members are also entitled to attorneys’ fees and costs of suit

COUNT V
Unlawful Agreement Between Google And Apple To Unreasonably Restrain Trade In Violation of
Sections 1 And 3 of the Sherman Act On Behalf of Plaintiff and The Class

243. Plaintiff incorporates by reference the preceding factual allegations.

244. In 2016, there was a possibility that Apple might wish to use Siri as a form of search

engine, a fact referenced in the 2018 e-mail from Braddi quoted above.

245. Google insisted that Apple shelve any such plan as a condition of extending its “Apple

Inc. Search Partnership” with Google. Apple acceded to this request and, in return, continued to

share Google’s advertising revenue through the use of Apple devices, as reflected in the $18 billion

payment it got in 2021. Google currently pays Apple 36% of the ad revenue obtained through Google

searches on Apple’s Safari web browser. This agreement was unlawful per se under Sections 1 and

3 of the Sherman Act (15 U.S.C. §§ 1, 3) or is otherwise an unreasonable restraint of trade under those

statutes.

246. Plaintiff and Class members paid excessive amounts for Google’s services that

would have been less but for this conduct and are entitled to receive treble damages.

247. Google’s “Apple Inc. Search Partnership” is ongoing and Plaintiff and Class

members are entitled to injunctive relief and other equitable remedies to prevent its continued

90
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 94 of 96

operation, given that they would otherwise have no adequate remedy at law.

248. Plaintiff and Class members are also entitled to attorneys’ fees and costs of suit.

X. RELIEF

A. DEMAND FOR JUDGMENT

Wherefore, Plaintiff requests that the Court enter judgment on their behalf and on behalf of

the Class defined herein, by ordering:

(a) This action may proceed as a class action, with Plaintiff serving as Class

Representative, and with Plaintiff’s counsel as Class Counsel;

(b) Judgment in favor of Plaintiff and members of the Class and against

Defendants;

(c) An award of statutory and other damages under 15 U.S.C. §15 for violations

of the antitrust laws by Defendants;

(d) Permanent injunctive relief pursuant to 15 U.S.C. §26, including, but not

limited to, guardrails to restore and ensure a fair and level competitive playing

field around Google’s publishing and dissemination of digital informational

content in its general search services and search advertising services, such

guardrails may include, among others, GAI products (including SGE and Bard),

which could include (but are not limited to): (i) revising Google’s practices so that

a Publisher who wishes to opt out of SGE would still show up on Google

searches; (ii) obtaining prior informed consent from Publishers whose website

data is used to train GAI; (iii) allowing rivals to access the training data Google

uses for chatbots such as Bard; (iv) modifying the payment terms Google

announced in 2023 in order to make them more favorable to Publishers; (v)

Google actively investing significant sums in supporting news or reference

91
 Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 95 of 96

dissemination by smaller Publishers; and (vi) establishing a monitoring

committee of neutral experts who would be charged with reporting annually on

Google’s compliance with the mandate of any injunction and on any potential

new anticompetitive conduct by it.

(e) Pre- and post-judgment interest on the damages awarded to Plaintiff and

members of the Class, and that such interest be awarded at the highest legal

rate from and after the date this class action complaint is first served on

Defendants;

(f) Defendants are to be jointly and severally responsible financially for the costs

and expenses of a Court approved notice program through post and media

designed to give immediate notification to the Class; and

(g) Further relief for Plaintiff and members of the Class as may be just and proper.

B. JURY DEMAND
Under Federal Rule of Civil Procedure 38(b), Plaintiff demands a trial by jury of all the claims

asserted in this Complaint so triable.

Dated: December 11, 2023 Respectfully submitted,

By: /s/ Michael D. Hausfeld

Michael D. Hausfeld
Scott A. Gilmore
Mandy Boltax
HAUSFELD LLP
888 16th Street N.W., Suite 300
Washington, DC 20006
Telephone: (202) 540-7200
[email protected]
[email protected]

92
Case 1:23-cv-03677-APM Document 1 Filed 12/11/23 Page 96 of 96

[email protected]

Scott Martin
HAUSFELD LLP
33 Whitehall Street, 14th Floor
New York, NY 10004
Telephone: (646) 357-1100
[email protected]

Michael P. Lehmann
HAUSFELD LLP
600 Montgomery Street, Suite 3200
San Francisco, CA 94111
Telephone: (415) 633-1908
[email protected]

Michael L. Roberts
Erich P. Schork
Kelly A. Rinehart
ROBERTS LAW FIRM US, PC
1920 McKinney Avenue, Suite 700
Dallas, TX 75204
Telephone: (501) 821-5575
[email protected]
[email protected]
[email protected]

93
 Case 1:23-cv-03677-APM Document 1-1 Filed 12/11/23 Page 1 of 2

8 ÿ7ÿÿ0ÿ
01213ÿ05267ÿ8966 ÿ
1ÿÿ31 18 66 8ÿ
+(/(1$:25/'&+521,&/(//& *22*/(//&DQG$/3+$%(7,1&
$PSKLWKHDWUH3DUNZD\
0RXQWDLQ9LHZ&$

ÿ !"#$ÿ%ÿ&'()*'"'ÿ%ÿ%)&(#ÿ+)(#'*ÿ,+-)"#)%%ÿ.....................ÿ  !"#$ÿ%ÿ&'()*'" 

1
'ÿ%ÿ%)&(#ÿ+)(#'*ÿ*'%'"*-"#ÿ...
ÿ0 8 
ÿ31  1 ÿ
0 8 6 8ÿ
5 31 ÿ
..................ÿ
6/06 ÿ1ÿ08ÿ31 1ÿ0868ÿ " #'2ÿ)"ÿ+-"*ÿ "*'3"-#) "ÿ-('(4ÿ!('ÿ#5'ÿ+ -#) "ÿ %ÿ#5'ÿ#&-#ÿ %ÿ+-"*ÿ)"6 +6'*ÿ
7-## &"'$(ÿ8%)&3ÿ"-3'4ÿ-**&'((4ÿ-"*ÿ#'+',5 "'ÿ"!39'&: -## &"'$(ÿ8)%ÿ;" <":ÿ
0LFKDHO'+DXVIHOG
+$86)(/'//3
WK6WUHHW1:6XLWH
:DVKLQJWRQ'&
 

11ÿ8=,818ÿ5ÿ071810 15 111ÿ01 1@6891ÿ5ÿ71013ÿ7 168ÿ8,+-'ÿ-"ÿ>ÿ)"ÿ"'ÿ9 ?ÿ% &ÿ

B Cÿÿ!D,MN(ODJÿEFLOPPGHIJKHJL B Sÿÿ8%!DHR(HDÿIEFNMÿTUGHIHJVKHLOFJJLÿ"FLÿNÿ,NILW:ÿ OLOZHJÿFPÿL[OVÿ(LNLHÿ BC BC )JYFI\FINLHRÿFIÿ,IOJYO\NMÿ,MNYHÿ BX BX
+-'ÿ-"ÿ>ÿ)"ÿ"'ÿ9 ?ÿ"+$:ÿ
[
,+-)"#)%%ÿ-"*ÿ"'ÿ9 ?ÿ% &ÿ*'%'"*-"#:ÿ57ÿ1267811ÿ0868ÿ531Aÿ
 ÿ  ÿ  ÿ  ÿ

B Qÿÿ!D*H(PDHÿJEFRNGJHLIJKHJL B Xÿ*O8)JGRHOIYVNOLLWHÿOLOZHJV[O\ÿFP OLOZHJÿFPÿ-JFL[HIÿ(LNLHÿ BQ BQ F)FJPPÿYÿ99FUUI\VVOOFJJIHHNVVLVVHÿÿROOJJÿNÿÿ-J#[JRÿOF,VLÿI[(OJHLYNIÿOL(\HLNÿNMÿL,HMÿNYHÿB` B`

ÿÿ,NILOHVÿOJÿOLHKÿ))):ÿ OLOZHJÿFIÿ(U]^HYLÿFPÿNÿ BS BS Ba Ba
%FIHO_JÿFUJLIWÿ %FIHO_Jÿ"NLOFJÿ
12086ÿ881bc6 ÿÿ 076ÿ5ÿ801
d7ÿeÿ/ÿfeÿgeÿ7higjklÿ lÿhmhÿnhÿjojnehnÿkgpjÿ0pnÿgqÿ7hfgeÿerÿgeÿfeÿÿ7gjjnogerfeiÿhpjÿgqÿ8pfhÿ
B ÿÿÿstuvuwxyu B =ÿÿÿÿÿ{|~wy†}wt~~‡uvÿ‡t|ÿ‚xwƒ„ÿ B 0ÿÿÿÿÿ”|s’‘vv|•ÿtvyuw~uv’|ÿs“|t‡ƒ
X B ÿÿÿÿšw
ÿ™|‘†}w~wƒÿ”|yuw~vtvt“ÿ
|w„{w|v‘vt~wƒÿ
; ÿehfhjpnhÿ ˆÿfjodeÿ ‰ÿcrf7jÿ7hÿ ÿÿt‚xt‡uv}tÿ
ˆ‰ÿfjodeÿjgrp7hÿ3ffdfhkÿ 8g7fdÿ87pjfhkÿ ekÿehpjÿgqÿnpfhÿqjg‹ÿekÿ7higjkÿ
ˆÿnnpdhlÿ3fdÿŠÿ8derjÿ Œÿ91ÿˆ–‰qqÿ ‹kÿÿnd7hrÿqgjÿhmfnÿ7higjkÿgqÿ
ˆˆÿrjdÿ6‹odgkjnÿ3ffdfhkÿ Œÿ=d7—ÿ3peiÿ–ˆÿ 7nÿnnfie‹ehÿÿ
ˆ ÿcjfeÿ Œˆÿ1˜01˜˜ÿ‰iÿ ›1qÿehfhjpnhlÿhmeÿÿigjen›ÿ
ˆ ‰ÿcjfeÿjgrp7hÿ3ffdfhkÿ Œ ÿ881ÿ fhdÿ/21ÿ
ˆ‰ÿcghgjÿ2mf7dÿ
ˆ‰‰ÿcghgjÿ2mf7dÿjgrp7hÿ3ffdfhkÿ 5hmjÿŒ8‰hÿ7hp8h1ÿnÿ‰iÿ
ˆŒÿ5hmjÿjngedÿ1epjkÿ –ÿijf7pdhpjdÿ7hnÿ
ˆŒÿcrf7dÿcdoj7hf7ÿ –ˆÿ6efjge‹ehdÿchhjnÿ
ˆŒ‰ÿjgrp7hÿ3ffdfhkÿ  –ÿ5hmjÿ8hhphgjkÿ7hfgenÿ1qÿ
ˆŒŽÿ9dhmÿ0jmj‹7phf7dÿÿ ÿÿr‹fefnhjhfÿie7kÿfnÿÿ
ÿÿÿÿÿÿÿjngedÿ1epjkÿjgrp7hÿ3ffdfhkÿ ÿÿ1egdrÿ
ˆŒÿnnhgnÿjgrp7hÿ3ffdfhkÿ
B7dÿ6
[ ÿÿÿœ|t|w~ÿv’vÿžšuŸ|w
jgojhkÿ =e—jpoh7kÿ
ÿÿÿÿÿ57ÿ B ÿrÿÿ{jwd}ÿÿ¡¥|ÿÿ8œ|pfhntÿ|w~ÿv’v
ÿ3erÿ0ger‹ehfgeÿ ÿoodÿ¢ÿ080ÿ‰ÿ Žÿ¥nÿ08ÿodfehfqqÿgjÿÿ Œ‰ÿ5hmjÿ1‹‹fijhfgeÿ7hfgenÿ
ÿgj7dgnpjÿ ˆÿ˜fhmrj£dÿÿ080ÿ‰Žÿ ÿÿÿÿÿÿÿrqerehÿ Žÿ77—hjÿ1eqdpe7rÿÿ
ˆÿ7ehlÿ3nÿŠÿ67h‹ehÿ Žÿ178 mfjrÿjhkÿŒÿ080ÿ ÿÿÿÿÿÿÿŠÿ0gjjpohÿ5jief¦hfgeÿ
 ÿgjhnÿhgÿ3erÿ j
fnge  j ÿ h
fh
fgenÿ ÿÿŽŒ–ÿ ÿ0genp‹jÿ0jrfhÿ
 ‰ÿgjhÿjgrp7hÿ3ffdfhkÿ ‰ˆ‰ÿhmÿedhkÿ gjqfhpjedhkÿ ‰ÿdomgeÿ0genp‹jÿÿ
–ÿddÿ5hmjÿ7dÿjgojhkÿ ‰ ÿcer‹pnÿŠÿ5hmjÿ Œ‰ÿjpiÿ7dhrÿ8f¦pjÿgqÿ ÿÿÿÿÿÿÿjgh7hfgeÿ7hÿ0¨ÿ
‰‰ÿ0ffdÿ7fimhnÿ ÿÿÿÿÿÿÿjgojhkÿÿ080ÿÿ –ÿ0d8hddfhÿ2ÿ
jngedÿjgojhkÿ ‰ ‰ ‰ ÿj f
ngeÿ0g erf
hfgen
ÿ Œ–ÿ5hmjÿ ‰ÿ87pjfhfn0g‹‹grfhfnÿ
ˆŽÿ5hmjÿjprÿ ‰Œÿ0ffdÿhfeÿ¤ÿ0gerfhfgenÿ ÿÿÿÿÿÿÿ6¥7meiÿ
ˆŽÿjphmÿfeÿ3erfeiÿ ÿÿgqÿ0geqfe‹ehÿ 5hmjÿ8hhphnÿ –Œÿjfhjhfgeÿ
ˆÿ5hmjÿjngedÿjgojhkÿ ˆŽ‰ÿdnÿ0df‹nÿ7hÿ ––ÿr‹fefnhjhfÿjg7rpjÿÿ
ÿÿÿÿÿÿÿ‹iÿ j
g o jh k ÿ
7f imhn
ÿ ˆŽŒÿ§pfÿ‹ÿˆÿ080ÿ ÿÿ7h7f£ÿgjÿoodÿgqÿ
ˆ‰ÿjgojhkÿ‹iÿÿ ÿ0gokjfimhnÿ ˆŽ–ÿ ÿÿÿÿÿÿÿie7kÿ7fnfgeÿ
ÿÿjgrp7hÿ3ffdfhkÿ ˆÿhehÿ ÿ8hhÿ7oogjhfge‹ehÿ –‰ÿ0genhfhphfgedfhkÿgqÿ8hhÿ
ˆ‰ÿhehÿ¤ÿjfhrÿ£ÿ ˆÿ=e—nÿŠÿ=e—feiÿ ÿÿ8hhphnÿ
ÿÿÿÿÿÿÿjpiÿoodf7hfgeÿ ‰ÿ0g‹‹j7100ÿ7hnh7ÿÿ –ÿ5hmjÿ8hhphgjkÿ7hfgenÿ
 ÿjr‹j—ÿ Œ ÿogjhhfgeÿÿ ÿÿfqÿeghÿr‹fefnhjhfÿie7kÿ
ÿqerÿjrÿ87jhnÿ7hÿgqÿÿ Œ ÿhpjdf¦hfgeÿÿ ÿÿjf£ÿgjÿjf7kÿ7hÿ
ÿÿŒÿ 8ÿ ÿÿoodf7hfgeÿ
 Case 1:23-cv-03677-APM Document 1-1 Filed 12/11/23 Page 2 of 2
0ÿ23 ÿÿÿ567869ÿ 9ÿ
ÿÿÿÿÿÿÿÿ
0ÿ3ÿÿÿ/0120834ÿ
569760636463ÿ
0ÿGÿ3ÿÿÿJKLMN6O672ÿM74ÿ 0ÿUÿ3ÿÿÿV4W834ÿX 63ÿ
ÿ ÿ
++8ÿ&9&"ÿ:&;<%ÿÿ=>"?>!%ÿÿ
ÿ
PQÿRC>ÿHÿG!H>%&!ÿ-)%ÿ
ÿ
#8ÿ:)9?ÿHÿ.H"%Cÿÿ
ÿÿÿÿ2!"ÿÿ PQÿS%<ÿ*%%%?ÿ-)%&!ÿÿ
#ÿ$%&!'()%ÿ*!%!)ÿ ÿÿÿÿÿÿÿ@)&%&Aÿ)Bÿ;!C'DBÿÿ ÿÿÿÿÿÿÿ*%C!%ÿY!ÿ
+,ÿÿÿÿ-"&!ÿÿ ÿÿÿÿÿÿÿ!%&!"ÿ&;&!Bÿÿ ÿÿÿÿÿÿÿ@&HÿT&9)?ÿ-)%Eÿ ÿÿÿÿÿÿÿ@D)"C&!;ÿ9%!Eÿ
ÿÿÿÿÿÿÿ.%&!ÿ ÿÿÿÿÿÿÿC&)&>&!%&!BÿC&&"&%?Bÿ;Bÿÿ ÿ
ÿ ÿÿÿÿÿÿÿ"&;&!Bÿ%"&%&!Eÿ ÿ
ÿ ÿ ÿ
F@GHÿÿBÿ")%ÿ%<&ÿC)IEFÿ F@GHÿÿBÿ")%ÿ%<&ÿC)IEFÿ
0ÿZ3 ÿÿÿX67 /[LVMÿÿ
ÿÿÿÿÿÿÿ\33]80120834^ÿ
0ÿY3 ÿÿÿK4a8ÿ6O61ÿ[6ba49ÿ 0ÿ$3ÿÿÿ 34674ÿ
ÿÿÿÿÿÿÿ\33]80120834^ÿ ÿ
0ÿc3ÿÿÿg4aÿ88]hWb8ÿ
ÿ
_#ÿR&ÿYÿ*%!CCÿ-)%ÿ
ÿ
++#ÿ(%&!;ÿ@&Hÿ!%ÿ(%&!;ÿ:&;<%ÿÿ
##ÿG!!)ÿ
#8ÿ$&!ÿ
ÿ
++#ÿ&9&"ÿ:&;<%ÿÿ(%&!;ÿÿ
_8ÿY'$;>%3ÿ:"%&!ÿ ÿÿÿÿÿÿÿ-)%Eÿ #ÿ$&""ÿ-)%ÿ ÿÿÿÿÿÿÿ@&Hÿ(%&!;ÿ:&;<%ÿ-)%Eÿÿ
_+ÿYÿ:&"`?ÿ-)%ÿ ++ÿ&!;'-))>>C%&!ÿ #+ÿc;%&"ÿG!%>!%ÿ ÿ
_#ÿR>&"?ÿ!Cÿ$C&)"ÿÿ ++ÿS%<ÿ&9&"ÿ:&;<%ÿ #ÿ:)9?ÿHÿS9?>!%ÿÿÿÿÿÿ ÿ
ÿÿÿÿÿÿÿY9ÿ-)%ÿ ++ÿ->&)!ÿ`'.&&"&%&ÿÿÿ ÿÿÿÿÿÿÿdÿ=!H)>!%ÿHÿÿ ÿ
_QÿS%<ÿYÿY&%&;%&!ÿ ÿÿÿÿÿÿÿ=>"?>!%ÿÿ ÿÿÿÿÿÿÿUC;>!%ÿ ÿ
_Q#ÿ=>"3ÿ:%3ÿG!)3ÿ*)&%?ÿ-)%ÿ ++,ÿ->&)!ÿ`'.&&"&%&ÿÿÿ #ÿ:)9?ÿHÿS9?>!%ÿÿ ÿ
ÿÿÿÿÿÿÿS%<ÿ ÿÿÿÿÿÿÿHÿ(%!eÿf!H&%ÿ ÿ
++Pÿ=C)%&!ÿÿ #,ÿ*%)I<"Ceÿ*&%ÿ ÿ
ÿ #QÿS%<ÿ!%)%ÿÿ ÿ
#Qÿ!%)%ÿTC)%ÿY&&"&%?ÿ ÿ
#Q,ÿR!)<&ÿ ÿ
ÿ
(3ÿS:G2Gcÿ
ÿ

0ÿT#ÿS)&;&C!&!";ÿÿÿÿÿÿÿ0ÿÿÿÿÿ8ÿÿÿÿ:H>>ÿ*9%C%ÿÿÿÿ0ÿHÿ:>ÿ>-!C"C"ÿ%ÿ 0ÿ+ÿ:ÿ:&!%!%CCÿÿ0ÿHÿi>ÿ!!H%<Cÿÿ 0ÿ,Y&ÿ$%&;"%%&&jC!&ÿ%&)%ÿÿÿ0ÿ_.&ÿ-%&)%ÿU"ÿ%Cÿ;ÿÿ 0ÿY&Pÿ$%&;"%%&&jC!&ÿ%ÿ&)%ÿ

[
ÿÿÿÿÿÿÿ%ÿ %ÿ C&%&)%ÿ@)&H?Eÿÿ
H>ÿ$;3ÿ .&)%ÿR&"ÿ
UC;ÿ
(G3ÿ-k*=ÿSRÿ-iGScÿ@Gi=ÿi=ÿk3*3ÿG(GYÿ*i-iki=ÿkc.=:ÿlGÿmSkÿ-:=ÿRGYGc2ÿ-c.ÿl:Gi=ÿ-ÿf:G=Rÿ*i-i=$=ciÿSRÿ-k*=3Eÿ
ÿ

ÿ
86&6HFWLRQ7KHFRPSODLQWLVEURXJKWDJDLQVWGHIHQGDQWVIRUPRQRSROL]DWLRQDQGDWWHPSWHGPRQRSROL]DWLRQLQYLRODWLRQRI6HFWLRQ
ÿ ÿ ÿ
DFWVRIDFTXLVLWLRQDQGDGHIDFWRMRLQWYHQWXUHWKDWYLRODWH6HFWLRQDQDJUHHPHQWQRWWRFRPSHWHDQGW\LQJLQYLRODWLRQRI6HFWLRQVDQG
ÿ
(GG3ÿ:=nk=*i=.ÿGcÿ pqrpsÿtuÿvqtwÿtwÿxÿY-**ÿ .=$-c.ÿÿÿ!%LOOLRQ EHIRUHWUHEOLQJ p‚ƒ„ÿ†rw ÿ‡ˆ‰Šÿ‹ŒÿƒŽˆƒÿ‹ˆÿ„‡Ž‰‹ˆ‘ÿ
ÿÿÿÿÿÿÿÿS$TY-Gciÿ ; -iGScÿyz{r|ÿu}|}p}~}ÿ€ÿ ÿÿÿÿÿÿÿÿÿÿÿUk:mÿ.=$-c.Aÿÿ m=*
ÿ
ÿÿÿ
ÿÿÿ
[ÿ
ÿÿÿÿÿÿÿÿÿÿÿÿcSÿ
ÿ
(GGG3ÿ:=Y-i=.ÿ-*=@*Eÿ
ÿ
’wƒƒÿ‹ˆ“‘”•„‘‹‡ˆ–ÿ ÿ ÿ ÿ
tŒÿŠƒ“—ÿ‰ƒ“ƒÿ„‡Ž‰ƒ‘ƒÿ”ƒ‰‘ƒÿ„“ƒÿŒ‡”Žÿ
ÿÿÿÿÿÿÿÿÿÿGRÿ-cmÿ m=*ÿ [ cSÿÿ
ÿ ÿ
˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜ÿ *G2c-ik:=ÿSRÿ-iiS:c=mÿSRÿ:=S:.ÿ˜˜˜˜˜V0LFKDHO'+DXVIHOG
.-i=Aÿÿ˜˜ ˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜˜ÿ
ÿ
Gc*i:kiGSc*ÿRS:ÿS$TY=iGc2ÿG(GYÿS(=:ÿ*==iÿU*j++ÿ
-%<&%?ÿHÿ&9&"ÿ9ÿ*<%ÿ
ÿ
v‚ƒÿ™wš››ÿ„‹œ‹‰ÿ„‡œƒ”ÿ“‚ƒƒ‘ÿˆÿ‘‚ƒÿ‹ˆŒ‡”Ž‘‹‡ˆÿ„‡ˆ‘‹ˆƒÿ‚ƒ”ƒ‹ˆÿˆƒ‹‘‚ƒ”ÿ”ƒ‰„ƒ“ÿˆ‡”ÿ“•‰ƒŽƒˆ‘“ÿ‘‚ƒÿŒ‹‰‹ˆ“ÿˆÿ“ƒ”œ‹„ƒ“ÿ‡Œÿ‰ƒ‹ˆ“ÿ‡”ÿ‡‘‚ƒ”ÿƒ”“ÿ“ÿ”ƒž•‹”ƒÿ
ŸŠÿ‰ —ÿƒ¡„ƒ‘ÿ“ÿ”‡œ‹ƒÿŸŠÿ‰‡„‰ÿ”•‰ƒ“ÿ‡Œÿ„‡•”‘}ÿÿv‚‹“ÿŒ‡”Ž—ÿ”‡œƒÿŸŠÿ‘‚ƒÿ™•‹„‹‰ÿp‡ˆŒƒ”ƒˆ„ƒÿ‡Œÿ‘‚ƒÿyˆ‹‘ƒÿw‘‘ƒ“ÿ‹ˆÿwƒ‘ƒŽŸƒ”ÿ¢£¤›—ÿ‹“ÿ”ƒž•‹”ƒÿŒ‡”ÿ‘‚ƒÿ•“ƒÿ‡Œÿ‘‚ƒÿ
p‰ƒ”ÿ‡Œÿp‡•”‘ÿŒ‡”ÿ‘‚ƒÿ•”‡“ƒÿ‡Œÿ‹ˆ‹‘‹‘‹ˆÿ‘‚ƒÿ„‹œ‹‰ÿ‡„ƒ‘ÿ“‚ƒƒ‘}ÿÿp‡ˆ“ƒž•ƒˆ‘‰Š—ÿÿ„‹œ‹‰ÿ„‡œƒ”ÿ“‚ƒƒ‘ÿ‹“ÿ“•ŸŽ‹‘‘ƒÿ‘‡ÿ‘‚ƒÿp‰ƒ”ÿ‡Œÿp‡•”‘ÿŒ‡”ÿƒ„‚ÿ„‹œ‹‰ÿ„‡Ž‰‹ˆ‘ÿŒ‹‰ƒ}ÿÿ
¥‹“‘ƒÿŸƒ‰‡ ÿ”ƒÿ‘‹“ÿŒ‡”ÿ„‡Ž‰ƒ‘‹ˆÿ‘‚ƒÿ„‹œ‹‰ÿ„‡œƒ”ÿ“‚ƒƒ‘}ÿÿv‚ƒ“ƒÿ‘‹“ÿ„‡‹ˆ„‹ƒÿ‹‘‚ÿ‘‚ƒÿ|‡Žˆÿz•Žƒ”‰“ÿ‡ˆÿ‘‚ƒÿ„‡œƒ”ÿ“‚ƒƒ‘}ÿÿ
ÿ
G3ÿ p§yzv†ÿ§uÿ|rwt{rzprÿ§uÿut|wvÿ¥twvr{ÿ~¥xtzvtuu¨{rurz{xzvÿ’Ÿ–ÿp‡•ˆ‘Šÿ‡Œÿ”ƒ“‹ƒˆ„ƒ©ÿy“ƒÿ¢¢ªª¢ÿ‘‡ÿ‹ˆ‹„‘ƒÿ‰‹ˆ‘‹ŒŒÿ‹Œÿ”ƒ“‹ƒˆ‘ÿ
‡Œÿ«“‚‹ˆ‘‡ˆ—ÿ{p—ÿ¬¬¬¬¬ÿ‹Œÿ‰‹ˆ‘‹ŒŒÿ‹“ÿ”ƒ“‹ƒˆ‘ÿ‡Œÿyˆ‹‘ƒÿw‘‘ƒ“ÿŸ•‘ÿˆ‡‘ÿ«“‚‹ˆ‘‡ˆ—ÿ{p—ÿˆÿ£££££ÿ‹Œÿ‰‹ˆ‘‹ŒŒÿ‹“ÿ‡•‘“‹ƒÿ‘‚ƒÿyˆ‹‘ƒÿw‘‘ƒ“}ÿ
ÿ
GGG3ÿ ptvt­rzwqt~ÿ§uÿ~|tzpt~x¥ÿ~x|vtrw©ÿv‚‹“ÿ“ƒ„‘‹‡ˆÿ‹“ÿ„‡Ž‰ƒ‘ƒÿ‡ˆ‰Šÿ‹Œÿ‹œƒ”“‹‘Šÿ‡Œÿ„‹‘‹®ƒˆ“‚‹ÿ“ÿ“ƒ‰ƒ„‘ƒÿ“ÿ‘‚ƒÿ̄“‹“ÿ‡Œÿ™•”‹“‹„‘‹‡ˆÿ
•ˆƒ”ÿwƒ„‘‹‡ˆÿtt}ÿ
ÿ
G(3ÿ pxwrÿxwwt°z±rzvÿxz{ÿzxvy|rÿ§uÿwytv©ÿv‚ƒÿ““‹ˆŽƒˆ‘ÿ‡Œÿÿ²•ƒÿ‘‡ÿŠ‡•”ÿ„“ƒÿ‹‰‰ÿƒƒˆÿ‡ˆÿ‘‚ƒÿ„‘ƒ‡”ŠÿŠ‡•ÿ“ƒ‰ƒ„‘ÿ‘‚‘ÿŸƒ“‘ÿ
”ƒ”ƒ“ƒˆ‘“ÿ‘‚ƒÿ”‹Ž”Šÿ„•“ƒÿ‡Œÿ„‘‹‡ˆÿŒ‡•ˆÿ‹ˆÿŠ‡•”ÿ„‡Ž‰‹ˆ‘}ÿ†‡•ÿŽŠÿ“ƒ‰ƒ„‘ÿ‡ˆ‰Šÿ‡ˆƒÿ„‘ƒ‡”Š}ÿÿ†‡•ÿŽ•“‘ÿ‰“‡ÿ“ƒ‰ƒ„‘ÿ‡ˆƒÿ„‡””ƒ“‡ˆ‹ˆÿ
ˆ‘•”ƒÿ‡Œÿ“•‹‘ÿŒ‡•ˆÿ•ˆƒ”ÿ‘‚ƒÿ„‘ƒ‡”Šÿ‡Œÿ‘‚ƒÿ„“ƒ}ÿ
ÿ
(G3ÿ pxywrÿ§uÿxpvt§z©ÿp‹‘ƒÿ‘‚ƒÿy}w}ÿp‹œ‹‰ÿw‘‘•‘ƒÿ•ˆƒ”ÿ‚‹„‚ÿŠ‡•ÿ”ƒÿŒ‹‰‹ˆÿˆÿ”‹‘ƒÿÿŸ”‹ƒŒÿ“‘‘ƒŽƒˆ‘ÿ‡Œÿ‘‚ƒÿ”‹Ž”Šÿ„•“ƒ}ÿ
ÿ
(GGG3ÿ |r¥xvr{ÿpxwr’w–—ÿtuÿxz†©ÿtŒÿŠ‡•ÿ‹ˆ‹„‘ƒÿ‘‚‘ÿ‘‚ƒ”ƒÿ‹“ÿÿ”ƒ‰‘ƒÿ„“ƒ—ÿŠ‡•ÿŽ•“‘ÿ„‡Ž‰ƒ‘ƒÿÿ”ƒ‰‘ƒÿ„“ƒÿŒ‡”Ž—ÿ‚‹„‚ÿŽŠÿŸƒÿ‡Ÿ‘‹ˆƒÿŒ”‡Žÿ
‘‚ƒÿp‰ƒ”³“ÿ§ŒŒ‹„ƒ}ÿ
ÿ
¯ƒ„•“ƒÿ‡Œÿ‘‚ƒÿˆƒƒÿŒ‡”ÿ„„•”‘ƒÿˆÿ„‡Ž‰ƒ‘ƒÿ‹ˆŒ‡”Ž‘‹‡ˆ—ÿŠ‡•ÿ“‚‡•‰ÿƒˆ“•”ƒÿ‘‚ƒÿ„„•”„Šÿ‡Œÿ‘‚ƒÿ‹ˆŒ‡”Ž‘‹‡ˆÿ”‡œ‹ƒÿ”‹‡”ÿ‘‡ÿ“‹ˆ‹ˆÿ‘‚ƒÿŒ‡”Ž}ÿÿ
 Case 1:23-cv-03677-APM Document 1-2 Filed 12/11/23 Page 1 of 2

01ÿ334ÿ56789ÿ4 ÿÿ ÿÿÿÿÿ8ÿ0

 !"#ÿ"0"#$ÿ%ÿ&7!"6!"ÿ16"
ÿ''''''''''ÿ%ÿ$ÿ''''''''''ÿ



HELENA WORLD CHRONICLE, LLC 
()*+,-+../01 
89  8ÿ0ÿ9




GOOGLE LLC and ALPHABET INC.

23.3,4*,-/01 
5677895ÿ;9ÿ<ÿ=;>;?ÿ<=@;89
"Aÿ/23.3,4*,-B0ÿ,*D3ÿ*,4ÿ*44E3001 GOOGLE LLC
1600 Amphitheatre Parkway
Mountain View, CA 94043

0ÿFÿ&ÿG77ÿ$7HÿIÿJ9
K&ÿ ÿHJÿ$7%ÿ7%87ÿ$ÿ&ÿÿÿJÿ5ÿIÿ&7ÿHJÿJÿ%7787HÿÿLÿ%ÿ4ÿHJÿ$ÿJ
%7ÿ&7ÿ7Hÿ7ÿ%ÿÿ7Hÿ7ÿI7JMÿ%ÿÿ$$7%ÿ%ÿ7NJ77ÿ$ÿ&7ÿ7Hÿ7ÿH7%G7HÿÿO7H9ÿ69ÿ89
P9ÿ ÿ55ÿ%ÿ5ÿLÿJÿÿ7%87ÿÿ&7ÿN$$ÿÿF7%ÿÿ&7ÿ&7HÿNÿ%ÿÿÿH7%ÿ67ÿ ÿ$
&7ÿO7H7%ÿ67ÿ$ÿ8ÿP%7H%79ÿÿ"&7ÿF7%ÿ%ÿÿÿG7ÿ7%87Hÿÿ&7ÿN$$ÿ%ÿN$$Qÿ%7JM
F&7ÿ7ÿHÿHH%7ÿ%7A Michael D. Hausfeld
HAUSFELD LLP
888 16th Street, NW
Suite 300
Washington, DC 20006
!$ÿJÿ$ÿÿ%7NHMÿRHI7ÿGJÿH7$ÿFÿG7ÿ77%7HÿIÿJÿ$%ÿ&7ÿ%77$ÿH7H7Hÿÿ&7ÿN9ÿ
Sÿÿÿ$7ÿJ%ÿF7%ÿ%ÿÿF&ÿ&7ÿ%9

TUVWXTÿ2YÿZTW[T\]ÿZXW\^ÿ_`ÿZ_a\b
7A
[+c,*-dE3ÿe.ÿZ)3EfÿeEÿ23gd-hÿZ)3Ef
 Case 1:23-cv-03677-APM Document 1-2 Filed 12/11/23 Page 2 of 2

01ÿ334ÿ56789ÿ4 ÿÿÿÿÿ8ÿ0ÿ57ÿ
8ÿ0ÿ9
 !ÿ!ÿ#$%&'$
()*+,ÿ,./0+12ÿ,*1345ÿ210ÿ6.ÿ7+4.5ÿ8+0*ÿ0*.ÿ/1390ÿ324.,,ÿ9.:3+9.5ÿ6;ÿ<.5=ÿ>=ÿ?+@=ÿA=ÿBÿ(4CC
DEÿÿFGÿHIJKLÿNOÿPIQPRPQSJTÿJIQÿUPUTLVÿPOÿJIWX
YÿG7787Zÿ[\ÿ7ÿÿHQJULX 9
] ^ÿ_7G\ÿ7G87ZÿE7ÿÿÿE7ÿZ8ZÿÿH`TJaLX
ÿHQJULX bÿG
] ^ÿ7FÿE7ÿÿÿE7ÿZ8ZcÿG7Z77ÿGÿÿ_7ÿFÿ[Z7ÿYEÿHIJKLX
dÿÿ_7GÿFÿ[7ÿ7ÿZÿZG7ÿYEÿG7Z7ÿE7G7d
ÿHQJULX dÿZÿ7Zÿÿ_\ÿÿE7ÿZ8ZcÿÿeYÿZZG7bÿG
] ^ÿ7G87ZÿE7ÿÿÿHIJKLÿNOÿPIQPRPQSJTX dÿYEÿ
ÿZ77Zÿ[\ÿYÿÿ7_ÿ7G87ÿFÿ_G7ÿÿ[7EFÿFÿHIJKLÿNOÿNfgJIPhJUPNIX
ÿHQJULX bÿG
] ^ÿG7G7ZÿE7ÿÿ7i77Zÿ[77 bÿG
] 1E7GÿHj`LaPOWXk
9
l\ÿF77ÿG7ÿm FGÿG87ÿZÿm FGÿ7G87dÿFGÿÿÿFÿm tutt 9
^ÿZ7G7ÿZ7Gÿ_7\ÿFÿ_7GnG\ÿEÿEÿFGÿÿG79
o7p
qLfRLfrjÿjPgIJUSfL

sfPIULQÿIJKLÿJIQÿUPUTL

qLfRLfrjÿJQQfLjj
0ZZÿFGÿG7GZÿ7_7Zÿ7G87dÿ7p
 Case 1:23-cv-03677-APM Document 1-3 Filed 12/11/23 Page 1 of 2

01ÿ334ÿ56789ÿ4 ÿÿ ÿÿÿÿÿ8ÿ0

 !"#ÿ"0"#$ÿ%ÿ&7!"6!"ÿ16"
ÿ''''''''''ÿ%ÿ$ÿ''''''''''ÿ



HELENA WORLD CHRONICLE, LLC 
()*+,-+../01 
89  8ÿ0ÿ9




GOOGLE LLC and ALPHABET INC.

23.3,4*,-/01 
5677895ÿ;9ÿ<ÿ=;>;?ÿ<=@;89
"Aÿ/23.3,4*,-B0ÿ,*D3ÿ*,4ÿ*44E3001 ALPHABET INC.
1600 Amphitheatre Parkway
Mountain View, CA 94043

0ÿFÿ&ÿG77ÿ$7HÿIÿJ9
K&ÿ ÿHJÿ$7%ÿ7%87ÿ$ÿ&ÿÿÿJÿ5ÿIÿ&7ÿHJÿJÿ%7787HÿÿLÿ%ÿ4ÿHJÿ$ÿJ
%7ÿ&7ÿ7Hÿ7ÿ%ÿÿ7Hÿ7ÿI7JMÿ%ÿÿ$$7%ÿ%ÿ7NJ77ÿ$ÿ&7ÿ7Hÿ7ÿH7%G7HÿÿO7H9ÿ69ÿ89
P9ÿ ÿ55ÿ%ÿ5ÿLÿJÿÿ7%87ÿÿ&7ÿN$$ÿÿF7%ÿÿ&7ÿ&7HÿNÿ%ÿÿÿH7%ÿ67ÿ ÿ$
&7ÿO7H7%ÿ67ÿ$ÿ8ÿP%7H%79ÿÿ"&7ÿF7%ÿ%ÿÿÿG7ÿ7%87Hÿÿ&7ÿN$$ÿ%ÿN$$Qÿ%7JM
F&7ÿ7ÿHÿHH%7ÿ%7A Michael D. Hausfeld
HAUSFELD LLP
888 16th Street, NW
Suite 300
Washington, DC 20006
!$ÿJÿ$ÿÿ%7NHMÿRHI7ÿGJÿH7$ÿFÿG7ÿ77%7HÿIÿJÿ$%ÿ&7ÿ%77$ÿH7H7Hÿÿ&7ÿN9ÿ
Sÿÿÿ$7ÿJ%ÿF7%ÿ%ÿÿF&ÿ&7ÿ%9

TUVWXTÿ2YÿZTW[T\]ÿZXW\^ÿ_`ÿZ_a\b
7A
[+c,*-dE3ÿe.ÿZ)3EfÿeEÿ23gd-hÿZ)3Ef
 Case 1:23-cv-03677-APM Document 1-3 Filed 12/11/23 Page 2 of 2

01ÿ334ÿ56789ÿ4 ÿÿÿÿÿ8ÿ0ÿ57ÿ
8ÿ0ÿ9
 !ÿ!ÿ#$%&'$
()*+,ÿ,./0+12ÿ,*1345ÿ210ÿ6.ÿ7+4.5ÿ8+0*ÿ0*.ÿ/1390ÿ324.,,ÿ9.:3+9.5ÿ6;ÿ<.5=ÿ>=ÿ?+@=ÿA=ÿBÿ(4CC
DEÿÿFGÿHIJKLÿNOÿPIQPRPQSJTÿJIQÿUPUTLVÿPOÿJIWX
YÿG7787Zÿ[\ÿ7ÿÿHQJULX 9
] ^ÿ_7G\ÿ7G87ZÿE7ÿÿÿE7ÿZ8ZÿÿH`TJaLX
ÿHQJULX bÿG
] ^ÿ7FÿE7ÿÿÿE7ÿZ8ZcÿG7Z77ÿGÿÿ_7ÿFÿ[Z7ÿYEÿHIJKLX
dÿÿ_7GÿFÿ[7ÿ7ÿZÿZG7ÿYEÿG7Z7ÿE7G7d
ÿHQJULX dÿZÿ7Zÿÿ_\ÿÿE7ÿZ8ZcÿÿeYÿZZG7bÿG
] ^ÿ7G87ZÿE7ÿÿÿHIJKLÿNOÿPIQPRPQSJTX dÿYEÿ
ÿZ77Zÿ[\ÿYÿÿ7_ÿ7G87ÿFÿ_G7ÿÿ[7EFÿFÿHIJKLÿNOÿNfgJIPhJUPNIX
ÿHQJULX bÿG
] ^ÿG7G7ZÿE7ÿÿ7i77Zÿ[77 bÿG
] 1E7GÿHj`LaPOWXk
9
l\ÿF77ÿG7ÿm FGÿG87ÿZÿm FGÿ7G87dÿFGÿÿÿFÿm tutt 9
^ÿZ7G7ÿZ7Gÿ_7\ÿFÿ_7GnG\ÿEÿEÿFGÿÿG79
o7p
qLfRLfrjÿjPgIJUSfL

sfPIULQÿIJKLÿJIQÿUPUTL

qLfRLfrjÿJQQfLjj
0ZZÿFGÿG7GZÿ7_7Zÿ7G87dÿ7p