Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 1 of 26 PageID #:1

IN THE U.S. DISTRICT COURT

FOR THE NORTHERN DISTRICT OF ILLINOIS
EASTERN DIVISION

Jacqueline Huskey, on behalf of herself Case No.: 22-cv-7014

and all others similarly situated,
CLASS ACTION COMPLAINT
Plaintiff,
Jury Trial Requested
v.

State Farm Fire & Casualty Company,

Defendant.

CLASS ACTION COMPLAINT

1. This is a class action brought by Jacqueline Huskey (“Plaintiff”), on behalf of

herself and similarly situated homeowners insurance policyholders,1 against State Farm Fire &

Casualty Company (“Defendant” or “State Farm”) under the Fair Housing Act (“FHA”), 42 U.S.C.

§ 3601 et seq. Plaintiff seeks remedies for herself and the class for the discriminatory effects of

State Farm’s homeowners insurance claim processing policies and practices. Plaintiff alleges and

avers as follows:

2. State Farm does not treat its Black and white homeowners insurance policyholders

equally. A large-scale survey of about 800 white and Black homeowners with State Farm

homeowners insurance policies across the Midwest—defined here as Illinois, Indiana, Michigan,

1
For the purposes of this Complaint, the term “home” includes single-family detached
houses; single-family attached houses; condominiums or cooperatives; mobile or manufactured
homes; and any other type of dwelling that is covered by an HO-01, HO-02, HO-03, HO-05, HO-
06, HO-07, or HO-08 homeowners insurance policy. The term “homeowners insurance
policyholders” contemplates persons with insurance policies designed to cover losses and damages
to such properties when used as a personal private residence, and not for rental or other commercial
purposes.
1
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 2 of 26 PageID #:2

Missouri, Ohio, and Wisconsin—has revealed marked disparities between State Farm’s handling

of claims filed by white homeowners relative to claims filed by Black homeowners.

3. As a general matter, claims made by Black homeowners are disproportionately

subjected to greater suspicion than claims made by their white counterparts: State Farm is more

likely to request additional claims documentation from Black claimants and Black claimants must

have more interactions with State Farm employees to resolve their claims. Accordingly, State

Farm takes longer to process and approve claims made by Black customers than it does to process

and approve claims made by white customers. Black homeowners thus experience greater

inconvenience as well as detrimental impact to the value of their home and their quality of life as

necessary repairs remain unaddressed for months on end. These delays mean that Black State

Farm policyholders receive a less valuable product than white State Farm policyholders.

4. On information and belief, this disparity is caused by State Farm’s use of automated

claims processing methods and machine-learning algorithms (collectively, “claims processing

methods”) to screen out potentially fraudulent or complex (“high touch”) claims from legitimate

or straightforward (“low touch” or “no touch”) claims. State Farm’s claims processing methods

make predictions and decisions about whether a claim might be fraudulent, how much scrutiny it

requires, and how it should be processed, and in doing so, rely on (1) biometric data that function

as proxies for race, such as physical appearance, genetics, and voice; (2) intrusive behavioral data

that function as proxies for race, such as geolocation, social media presence, and browser search

history; and (3) historical housing and claims data that are themselves infected with racial bias.

On information and belief, State Farm’s claims processing methods unjustifiably identify claims

submitted by Black homeowners for additional scrutiny at higher rates than their white

counterparts. As a result, State Farm’s claims processing methods have a demonstrable and

2
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 3 of 26 PageID #:3

widespread discriminatory impact on Black homeowners making insurance claims, in violation of

the FHA.

5. The delays and extra administrative hurdles that Black homeowners have

experienced and are experiencing because of State Farm’s discriminatory claims processing

methods cause tangible financial harm to these groups, and are unreasonable, vexatious, and

humiliating. Accordingly, Plaintiff seeks damages as well as declaratory and injunctive relief.

I. JURISDICTION AND VENUE

6. This Court has jurisdiction over Plaintiff’s Fair Housing Act claim under 28 U.S.C.

§ 1331 and 42 U.S.C. § 3613(a).

7. The Court has general personal jurisdiction over State Farm because it is domiciled

in Illinois.

8. Venue is proper in this judicial district under 28 U.S.C. § 1391(b), because a

substantial part of the events or omissions giving rise to the claim occurred in this judicial district,

State Farm regularly conducts business in this District, and the named Plaintiff resides in this

District.

II. PARTIES

9. Jacqueline Huskey (“Huskey”) is a Black resident of Matteson, Illinois, where she

owns a single-family home. Huskey has had a homeowners insurance policy with State Farm on

her home in Matteson since approximately March of 2021. On June 12, 2021, the roof of Huskey’s

home was damaged by hail. She promptly filed a homeowners insurance claim with State Farm.

This was her first time ever filing such a claim with State Farm. Huskey did not hear back from

State Farm with respect to her claim until more than a month after filing, on July 22, 2021. On

August 11, 2021, State Farm sent an adjuster to inspect the damage, but the adjuster declined to

3
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 4 of 26 PageID #:4

conduct an inspection of the outside of the roof and only provided an estimate for the damage to

the inside of the home. Huskey had to call State Farm representatives several times before the

company agreed to send a third-party adjuster to inspect the outside of the roof. Nearly four

months after Huskey filed her claim, State Farm granted it, but only for the cost of internal repairs.

Huskey still has not reached a resolution with State Farm as to the damage to the outside of her

roof and has been unable to repair it. As a result of State Farm’s delay, Huskey experienced further

damage to her home—water damage to her kitchen and to two bathrooms caused by leaks in the

unrepaired roof—and a decrease to her home’s overall value. She estimates that she has had 20 to

30 conversations with State Farm regarding this claim, including as recently as May 2022.

10. Defendant State Farm Fire & Casualty Company is an Illinois corporation with its

principal office in Bloomington, Illinois. It issued homeowners insurance policies to Huskey and

sustained those policies during the statute of limitations period. It is a subsidiary of State Farm

Mutual Automobile Insurance Company, a large insurance company that sells homeowners

insurance policies in states across the country, including Illinois, Indiana, Michigan, Missouri,

Ohio, and Wisconsin. In 2021, State Farm’s total revenue was more than $82 billion. It is the

largest provider of homeowners insurance in Illinois, Indiana, Michigan, Missouri, and Ohio, and

the second largest in Wisconsin.

III. FACTS

11. Homeowners insurance is a prerequisite to home ownership, with banks and

mortgage companies generally requiring borrowers to acquire coverage as a precondition for

financing. It is also the vehicle by which homeowners protect the sizeable financial investment

they undertook when purchasing their home: If disaster strikes, a homeowner can rest easy

knowing that their policy will cover necessary repairs such that they will still have a place to live

4
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 5 of 26 PageID #:5

and that their home’s value will be preserved. In tandem, purchasing property and protecting it

with insurance can, over time, facilitate upward financial mobility and create generational wealth.

12. Though home ownership is a critical vehicle for building wealth in the United

States, it has long been a lopsided one: In 2020, 73.7% of white families owned homes, compared

to only 44% of Black families. This gap, driven by a host of factors, is exacerbated by existing

administrative barriers that can be both humiliating and financially onerous. Indeed, in 2021,

economists at Freddie Mac found that 12.5% of homes in majority Black areas are appraised below

the price agreed upon by the buyer and seller, compared to just 7.4% of homes in majority-white

areas. In recent years, researchers have also determined that relative to their home value, Black

homeowners pay higher mortgage rates at origination and post-origination of the loan, spend more

on mortgage insurance, face a higher share of maintenance costs, and pay more in property taxes

than white homeowners, resulting in a higher homeownership cost burden. These higher costs,

borne by Black homeowners, leave them more reliant than white homeowners on their

homeowners insurance policy when property damage occurs.

13. Unfortunately, race-based economic inequities currently manifest in the

homeowners insurance market in several ways, from how policies are sold, to how clients are

treated once they have coverage, to how claims on those policies are received and ultimately

adjudicated. Specifically, as to Black policyholders, State Farm’s claims processing methods

exacerbate existing barriers to achieving and maintaining home ownership, in violation of the

FHA.

A. Modern Race Discrimination and Algorithmic Bias

14. Despite the FHA’s prohibition on racial discrimination in the housing context, such

discrimination against Black homeowners has persisted for decades since the Act’s passage.

Racial disparities created by “redlining”—the systematic denial of financial services to residents

5
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 6 of 26 PageID #:6

of certain areas based on race—still exist today. One recent study estimates that, over the last 40

years, the typical homeowner in a redlined neighborhood experiences 52% less of the property

value increase realized by the typical homeowner in a “greenlined” neighborhood.

15. In the insurance industry, as elsewhere in our society, racial discrimination has

shifted from overt to covert. Even though race-based redlining is now illegal, discrimination has

persisted through practices such as using credit-based insurance scores,2 and discriminatory

underwriting guidelines that use age and home value as a proxy for race. Inequitable practices

such as these make it more difficult for Black homeowners to build wealth through homeownership

at the same rate as white homeowners.

16. Today, discrimination is perpetuated by the modern trend toward automation and

data mining. Institutions like insurance companies use algorithmic models to quickly analyze vast

troves of publicly available information (“data mining”) to detect patterns and assist in making

future decisions (“data analytics”). So-called “machine-learning” algorithms are expressly

designed to learn based upon the algorithm’s access to a designated data set or an algorithm-driven

search for data residing on the internet or in a confined database.

17. Unfortunately, algorithms too often have discriminatory effects, even where

demographic data, such as race, are not included as inputs. This is because algorithms can “learn”

to use omitted demographic features by combining other inputs that are correlated with race (or

another protected classification), like zip code, college attended, and membership in certain

groups. To illustrate, Amazon famously abandoned a facially neutral hiring algorithm in 2017

2
According to a 2022 study by the Urban Institute, “Majority-Black communities and
majority–Native American communities have the lowest median credit scores and the highest debt-
in-collection rates, subprime credit score rates, and use of high-cost payday and other nonbank
loans. These racial disparities reflect historical inequities that reduced wealth and limited
economic choices for communities of color.”
6
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 7 of 26 PageID #:7

because of its disparate impact on female candidates. There, the training data presented to the

algorithm consisted of resumes submitted to Amazon by applicants over a 10-year period, without

presenting data to the algorithm explicitly indicating the applicants’ gender. But most of these

applicants were white males. Rather than sort candidates by qualifications or merit, the algorithm

learned to favor male candidates by prioritizing language more commonly used by males,

penalizing the word “women’s” in resumes, and devaluing candidates who had graduated from all-

women’s colleges.

18. Algorithmic decision-making and data analytics are not, and should not be assumed

to be, race nor gender neutral. Too often, they reinforce and even exacerbate historical and existing

discrimination. Academics and government actors alike have cautioned that when approached

without appropriate forethought and oversight, data analytics “can reproduce existing patterns of

discrimination, inherit the prejudice of prior decision makers, or simply reflect the widespread

biases that persist in society. It can even have the perverse result of exacerbating existing

inequalities by suggesting that historically disadvantaged groups actually deserve less favorable

treatment.” Indeed, according to Federal Trade Center (“FTC”) Commissioner Kelly Slaughter,

“[i]n recent years, algorithmic decision-making has produced biased, discriminatory, and

otherwise problematic outcomes in some of the most important areas of the American

economy. . . . These harms are often felt most acutely by historically disadvantaged populations,

especially Black Americans and other communities of color.”

19. As described in more detail below, State Farm’s claims processing methods use

algorithmic decision-making that disproportionately subjects the claims of Black policyholders to

greater suspicion—and thereby greater administrative process and delay. In this way, State Farm

is reproducing and exacerbating existing patterns of race discrimination.

7
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 8 of 26 PageID #:8

B. State Farm’s Homeowners Insurance Claims Process Has a Disparate Impact

on Black Policyholders
20. In 2021, YouGov—a reputable online polling provider that employs industry best

practices and an internal quality assurance process—surveyed about 800 white or Black

homeowners with State Farm homeowners insurance policies across the Midwest3 to assess

whether there were racial disparities in State Farm’s homeowners insurance claim submission and

adjudication process. Among other data points, the survey collected data on where the

homeowners were located; how long their claims process took; whether they were asked to submit

additional paperwork after making a claim; how many interactions with a claims handler were

required to resolve the claim; and the ultimate outcome of their claim submission.

21. The survey of 648 white State Farm policyholders and 151 Black State Farm

policyholders showed large and statistically significant racial disparities between Black and white

homeowners regarding 1) the time it took for claims to be paid out by State Farm; 2) the

supplemental paperwork required as part of the claim adjudication process; and 3) the number of

interactions claimants had with State Farm employees prior to having their claims paid out.

22. Thirty-nine percent of white State Farm policyholder respondents had their claim

paid out in one month or less from time of submission; by contrast, only 30% of Black homeowner

respondents were paid out at the same rate. That means white homeowners were almost a third

more likely than Black homeowners to have their claim processed expeditiously (in less than

a month). This racial disparity is statistically significant; it would occur less than 5% of the time

as a result of random chance.

3
The survey defined the “Midwest” as Illinois, Indiana, Michigan, Missouri, Ohio, and
Wisconsin.
8
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 9 of 26 PageID #:9

23. The survey results also show a disparity in the administrative burdens imposed on

white and Black homeowners in the Midwest. Only 46% of white State Farm policyholder

respondents were asked to procure additional materials beyond those required to initiate the claims

submission process. By contrast, 64% of Black State Farm policyholder respondents were asked

to gather such materials. Thus, Black policyholders were 39% more likely to have to submit

extra paperwork to justify their claims. Again, this racial disparity is statistically significant; it

would occur less than 1% of the time as a result of random chance.

24. The survey also identified a statistically significant disparity by race for how many

interactions with State Farm employees were required of homeowners while their claim was being

processed. Some 51% of white policyholder respondents resolved their claims after only one-to-

three interactions with State Farm employees, contrasted with only 42% of Black respondents.

Only 49% of white respondents reported needing three or more interactions with someone at State

Farm to resolve their claim, contrasted with 58% of Black respondents—meaning that Black

policyholders were about 20% more likely than white policyholders to require three or more

interactions with State Farm before claim resolution. This racial disparity would occur less

than 10% of the time as a result of random chance.

25. Thus, Black State Farm policyholders disproportionately experience administrative

burdens; they have to put in more work and have more conversations, all while under greater

suspicion, just to get the same pay out as their white neighbors. And these results represent only

a portion of the true harm. It is reasonable to assume that some Black policyholders simply give

up on pursuing their claims due to the frustration of these added burdens.

9
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 10 of 26 PageID #:10

26. On information and belief, this race-based burden is caused by State Farm’s use of

discriminatory claims processing methods that disproportionately subject the claims filed by Black

homeowners to greater scrutiny than those filed by white homeowners, as described further below.

C. State Farm’s Automation of Claims Processing via Data Mining and Machine-
Learning Algorithms

27. The claims processing aspect of the insurance industry is notoriously opaque, with

little oversight or regulation geared toward increasing transparency for consumers and regulators

alike. Insurance companies such as State Farm do not typically publish their claims processing

guidelines, their claims processing algorithms, their claims outcomes, or other related data. What

the industry readily confirms, however, are the basics of the traditional home insurance claim

process: A homeowner who has suffered a loss submits a claim; the insurance company assigns

an adjuster to assess the loss and apply the wording of the homeowner’s policy to the adjuster’s

interpretation of the facts of the case. The adjuster can choose whether to request more

information, go to the scene, request other evaluations, or meet with witnesses.

28. Increasingly, however, State Farm and other insurers are automating claims

processing, or major aspects of it, and replacing human judgment with algorithms. A significant

number of claims are now handled without conversations with the homeowner or follow-up

investigation. State Farm instead relies on data analytics and machine-learning algorithms to

process the claims it receives. These tools are designed to, among other things, predict the

likelihood of fraud and to sort “no touch” or “low touch” claims (which are paid out immediately

or near immediately) from “high touch” claims (which trigger additional scrutiny).

29. State Farm does its best to keep the nature of its claims processing methods

confidential, asserting that disclosing any details publicly would undermine anti-fraud efforts. But

those disclosures that have been made—by State Farm and its third-party vendors—reveal that

10
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 11 of 26 PageID #:11

State Farm harnesses a variety of tools to collect extensive data about policyholders, and to use

that data in claims processing and fraud detection. It describes itself as at the “cutting edge of

analytics” and it actively recruits and hires employees with a background in data analytics, to “turn

data into actionable insights by leveraging a combination of Natural Language Processing,

Machine Learning, Artificial Intelligence, or other data science tools and concepts.”

30. The nature and quantity of personal consumer information that State Farm

cultivates is staggering. According to its website, State Farm collects data and information on its

policyholders including: classifications such as race, sex, marital status, familial status, and

gender; physical characteristics and/or descriptions; education, employment, employment history,

professional licenses or designations; financial information, medical information or health

insurance information; personal property records, products or services purchased, obtained, or

considered, or other purchasing or consuming histories or tendencies; biometric information such

as genetic, physiological, behavioral and biological characteristics that can be used to establish

individual identity, including but not limited to fingerprints, voiceprints, retina scans, and sleep,

health or exercise data; internet usage information such as browsing history, search history, and

information regarding customers’ interaction with a website, application, or advertisement;

geolocation data such as precise physical location or movements and travel patterns; and sensory

data such as audio recordings of customer care calls.

31. State Farm uses this data in a variety of ways, including “creating a profile about a

consumer reflecting the consumer’s preferences, characteristics, psychological trends,

predispositions, behavior, attitudes, intelligence, abilities and aptitudes,” and for a variety of

purposes, including processing claims and “protect[ing] against fraud.”

11
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 12 of 26 PageID #:12

32. On information and belief, State Farm uses a combination of internal and third-

party tools to leverage the vast troves of data it collects to process homeowners insurance claims.

Data regarding policyholders and their claims are stored, managed, and accessed primarily through

State Farm’s Enterprise Claim System (“ECS”), a proprietary web-based system used by State

Farm claims associates. ECS operates as an electronic file where information about a claim may

be found and updated in real time. In addition to providing a web-based interface for the entry and

storage of claim data, State Farm executes automated processes based on code and data

relationships. As an example, State Farm has disclosed that it licenses a third-party vendor’s

proprietary system, Technology Analytics for Claims (“TAC”), which uses text-based queries of

claims data from ECS to help detect and identify claims that might be fraudulent.

33. In fact, State Farm maintains relationships with a variety of third-party vendors that

offer software, integrations, and applications focused on insurance claims automation. Especially

relevant here, State Farm has a relationship with Duck Creek Technologies, an insurance-specific

software and analytics company that provides comprehensive claims management and fraud-

detection tools.

34. According to Duck Creek’s SEC filings, State Farm was its single largest customer

as of fiscal year 2019. On information and belief, State Farm continues to be one of Duck Creek’s

largest customers.

35. Per its website, Duck Creek offers a “comprehensive claims management solution

that helps insurers manage the entire claims lifecycle—from first notice of loss to settlement—in

a single integrated solution,” called Duck Creek Claims. Duck Creek Claims provides “end-to-

end claims workflows that enable high-touch to no-touch claim handling.” The system is

comprised of “dynamically-guided workflows, rule-driven automation, personalized user

12
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 13 of 26 PageID #:13

interfaces and data enrichment to automate processes where needed.” In practice, this means that

on receipt of a claim, the system uses predictive modeling or rules-based decision making to

determine how a claim should be treated and then structures the claim workflow accordingly. The

system’s view of appropriate workflow is impacted by claim type, but also by processing rules

(e.g., rules that determine whether a claim is high touch, low touch, or no touch) and by predictive

fraud modeling.

36. Duck Creek relies on third-party integrations to detect fraud and structure claims

workflow for customers like State Farm.

37. Since at least 2018, Duck Creek has partnered with FRISS, a “provider of AI-

powered risk and fraud detection solutions.”4 FRISS software integrates with Duck Creek’s

software to help insurers segment and route claims, based on their risk for fraud. FRISS does this

by conducting a real-time fraud assessment on every claim and assigning a risk score, known as

the “FRISS Score.” FRISS claims that its approach to fraud detection is unique in the ways in

which it incorporates several data elements into its fraud models: In order to generate the FRISS

Score, FRISS aggregates data internal to an insurer, including historical claims and policy data,

with data from external sources, and runs that data through a “hybrid model engine” composed of

four elements: (1) an insurer-specific set of business processing rules; (2) FRISS’s AI model

(based on thousands of decision-trees regarding fraud); (3) an AI text mining tool (designed to

search for questionable words and phrases and search for suspect language patterns); and (4) a

network analysis tool (designed to reveal connections and patterns among claims, people,

businesses, etc.). In its marketing materials, FRISS repeatedly promotes the fraud detection

4
Guidewire Software, which State Farm partnered with prior to Duck Creek Technologies
(as of at least 2017, if not before), also partners with FRISS and uses the FRISS Score.
13
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 14 of 26 PageID #:14

benefits to insurers from incorporating external data sources. According to FRISS, such sources

include “demographic data about the neighborhood, such as the degree of urbanization” (emphasis

added) or crime statistics and data harvested from social media crawling.

38. As explained in an industry presentation on FRISS: “Much like the way banks rely

on credit scores to assess the risk in lending to an individual, insurance carriers . . . rely on the

FRISS Score to assess the likelihood of fraudulent activity in claims.” A claim with a lower FRISS

score is deemed less risky than a claim with a higher risk score. A score of 50 or less, for example,

indicates that FRISS has little to no concerns about the claim, and can be quickly reviewed by

claims handlers or fast tracked to payment. A score slightly above 50 might indicate there are

some anomalies with the claim, and that a claims adjuster should spend a little extra time reviewing

these claims and perhaps forward them to a supervisor for a decision before payment can be made.

A high FRISS score suggests fraud is likely and the claim should be referred to triage or specific

claims adjusters for analysis.

39. According to Duck Creek and FRISS, “[a] FRISS Score is generated at every step

in the claims process: [first notification of loss], claim contact added, loss details added/changed,

before issuance of payment – as well as at any time on demand.” The software is then capable of

auto-generating tasks for the claims adjuster based on a claim’s FRISS Score. As explained in the

same industry presentation discussed above, the system is designed such that “honest customers

receive lightning-quick responses, while questionable claims are guided through further

processing with actionable insights and automated workflows” in the insurer’s claims management

system.

40. On information and belief, throughout the Class Period, State Farm has used Duck

Creek Claims, the FRISS Score, and/or other similar internal and external tools to process its

14
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 15 of 26 PageID #:15

homeowners insurance claims, determine how much scrutiny they deserve, and assign tasks to

claims handlers. All homeowners claims that State Farm receives are subject to processing via

these automated tools, including the claims of Plaintiff and the Class. On information and belief,

these tools rely on biased historical data and troves of suspect personal consumer data to

disproportionately and unjustifiably assign higher risk/suspicion scores to Black claimants,

resulting in additional scrutiny and delay in resolution of their claims.

41. In a May 2021 letter to the National Association of Insurance Commissioners

Special Committee on Race and Insurance, the Center for Economic Justice specifically cited

FRISS as a “new entrant[]” that is deploying “AI systems using data sources known to be biased

against communities of color” as part of insurers’ fraud detection.

42. In the same letter, the Center for Economic Justice noted that “no attention has been

given to potential racial bias in claims settlement and anti-fraud, despite the fact that these parts of

the insurance operation utilize big data and AI as much or more than [they do] for pricing.”

43. State Farm is registered with, frequently communicates with, and actively

participates in the programs of the National Association of Insurance Commissioners. On

information and belief, State Farm became aware of this letter shortly after it was sent but

continued to use FRISS and other claims automation tools.

44. A wealth of literature discusses the potential for bias resulting from algorithmic

decision-making. As the FTC has acknowledged, algorithmic bias is everywhere. Mounting

evidence reveals that algorithmic decisions can produce biased, discriminatory, and unfair

outcomes in a variety of high-stakes economic spheres including employment, credit, health care,

and housing. Antifraud algorithms are particularly susceptible to racial bias. As the Center for

Economic Justice adeptly explained to the National Association of Insurance Commissioners:

15
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 16 of 26 PageID #:16

“Biased antifraud algorithms become self-[fulfilling]—if there is racial bias in the claims you

identify as potential[ly] fraudulent and investigate, there will be racial bias in the claims identified

as fraudulent. You can’t find fraud in a claim you don’t investigate.” Thus, where biased claims

processing algorithms subject Black claimants to greater scrutiny, more fraud will be found among

Black claimants, resulting in continuing and increasing scrutiny of Black claimants.

45. On information and belief, State Farm also relies on natural language processing

(or “voice and text analytics”) via tools like TAC to process homeowners insurance claims.

Natural language processing is a branch of computer science concerned with giving computers the

ability to understand text and speech the way that human beings can. Academic research indicates

that natural language processing results in racial bias. For example, a comprehensive study by

Stanford University researchers published in the Proceedings of the National Academy of Sciences

in 2020 found “large racial disparities” in voice analytics, even when Black and white speakers

spoke “identical phrases.” According to the researchers, the “results point to hurdles faced by

African Americans in using increasingly widespread tools driven by speech recognition

technology.” Indeed, the researchers noted that “[t]hese disparities may . . . actively harm African

American communities when, for example, speech recognition software is used by employers to

automatically evaluate candidate interviews.” By the same logic, the use of voice analytics by an

insurer is likely to cause longer delays in processing claims for Black homeowners and/or greater

risk of fraud investigations.

46. In addition, a 2021 report from The Brookings Institution’s Artificial Intelligence

and Emerging Technology Initiative outlined how the underlying data in text analytics creates bias

in text analytics software:

Studying biases in widely used word embeddings trained on a corpus of 800 billion
words collected from the web reveals that names of African Americans tend to co-

16
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 17 of 26 PageID #:17

occur with unpleasant words. Measuring the relative association of names of

African Americans vs. names of white people with pleasant and unpleasant words
shows that the word embeddings contain negative associations for the concept of
an African American social group due to the biased depiction of the group on the
internet.

When these stereotypical associations propagate to downstream applications that

present information on the internet or make consequential decisions about
individuals, they disadvantage minority and underrepresented group members.

47. Any reasonable effort to examine State Farm’s claims processing methods for

disparate impact would have detected the statistically significant discrepancies described above.

This is particularly true because State Farm has access to all of the relevant data via its ECS and

thus can tell with precision how race affects its algorithms’ outputs. Thus, State Farm either does

not review its claim processing methods for bias or, having reviewed them, has refused to use less

discriminatory alternatives.

48. State Farm has acknowledged the risk of bias in its machine learning processes, but

only in the underwriting context: In its application for U.S. Patent No. 11,315,191 (issued April

26, 2022), State Farm states that machine learning models could be trained to include bias if bias

is present in the data sets used for training. The patent proposes identifying “undesired factors” to

control for “undesired prejudice or discrimination,” by teaching the machine learning to not

consider those factors in setting a premium amount. No publicly available information indicates

that State Farm has taught its technology to limit discriminatory outcomes in homeowners

insurance claims processing.

D. State Farm’s Automated Claims Processing Methods Disproportionately

Impact Black Policyholders

49. On information and belief, State Farm’s claims processing methods rely on

variables that are highly predictive of or are direct proxies for race and are trained on historically

biased housing and claims data. These variables impact State Farm’s algorithmic calculations and

17
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 18 of 26 PageID #:18

predictions as to how much scrutiny a claim deserves and, thereby, the workflow an assigned

claims handler should follow in processing the claim. As a result, State Farm’s claims processing

methods have subjected Black homeowners to longer wait times during which they frequently are

forced to endure substandard living situations for unreasonably extended periods of time. This, in

turn, results in embarrassment and humiliation for Black homeowners—not to mention the

extended devaluation of their homes. And the additional labor that Black homeowners must

expend to have their claims processed results in inconvenience and loss of time which could

otherwise be put to valuable use.

50. State Farm’s claims processing methods disproportionately and adversely affect

Black homeowners relative to white homeowners and cause Black policyholders to expend

comparatively more effort over a longer period than white policyholders to get their claims

resolved.

51. The challenges presented by State Farm’s claims processing methods produce two

distinct harms. First, Black policyholders have to endure substandard living conditions for longer

and suffer the humiliation and added expense of undue delay and administrative burden. Plaintiff

Jacqueline Huskey, for example, was forced to live with a defective roof and water damage to the

interior of her home as her exterior roof claim has languished with State Farm. She has spent hours

of her time diligently trying—yet inexplicably failing—to get her claim fully addressed. Second,

Black State Farm policyholders receive a less valuable insurance product for the same price as

white State Farm policyholders because it takes Black policyholders longer and more effort to

obtain repayment.

52. Considered from either vantage, the survey results discussed above substantiate the

disparate impact of State Farm’s discriminatory claims processing methods.

18
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 19 of 26 PageID #:19

53. State Farm has no legitimate business need to use claims processing methods that

rely upon data that include race and/or proxies for race and lead to discriminatory effects. State

Farm could deploy alternative claims processing methods that would eliminate or greatly reduce

discriminatory impact.

54. Indeed, it would be straightforward and low cost for State Farm to prevent its

algorithms from creating the disparate impact they do. Testing for bias, either by census tract or

by inferred demographics, is a standard approach supported by regulators and academic research.

Assessing an algorithms’ output to examine whether it varies by race is trivial. Had State Farm

taken even basic measures to identify bias, simple measures could have been taken to remediate

the bias. State Farm itself has acknowledged in its patent applications that it is able to identify

“undesired factors” to control for “undesired prejudice or discrimination,” by teaching the machine

learning to not consider those factors in setting a premium amount.

55. Plaintiff and the proposed Class continue to suffer harm because of State Farm’s

discriminatory acts, practices and policies, losing wealth from delayed payment, as well as from

the additional time and paperwork required from them to process their claims. Additionally, as

claims of Black homeowners drag on—with repairs not timely paid for or timely completed—the

value of Black homes disproportionately decrease relative to the value of white homes.

Unaddressed repairs can lead to other issues in the home: health and safety issues, citations for

code violations, and even condemnation of a property. Finally, as claim delays persist, Black

homeowners may arbitrarily and unreasonably be forced to live in substandard housing for

extended periods of time. Studies have shown that living in substandard housing can impact both

physical and mental health of homeowners and their families.

19
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 20 of 26 PageID #:20

56. On information and belief, State Farm’s current claim processing methods have

been in place since at least 2017, if not before, and have existed on a continuous basis since then.

These practices have resulted in systemic and continuing racial discrimination in the processing of

homeowners insurance claims in every year since 2017.

IV. CLASS ALLEGATIONS

57. Plaintiff repeats and re-alleges every allegation above as if set forth herein in full.

58. Plaintiff sues on her own behalf under Rules 23(a), (b)(2), (b)(3), and (c)(4) of the

Federal Rules of Civil Procedure.

59. Plaintiff seeks to bring a class action pursuant to the FHA on behalf of herself and

a class of similarly situated individuals defined as follows:

All Black individuals who maintained State Farm homeowners insurance policies at some
time during the Class Period covering property in Illinois, Indiana, Michigan, Missouri,
Ohio, and/or Wisconsin; who made a claim under such policy during the Class Period; and
who were required to submit additional paperwork or information (beyond submission of
the initial claim), had three or more interactions with State Farm claims handlers before
claim resolution, and/or waited more than a month for claim resolution.
For purposes of this Class Definition, the term “homeowners insurance policy” is defined
to include insurance policies designed to cover losses and damages to properties used as a
private personal residence, and not for rental or other commercial purposes. Such
properties include single-family detached houses; single-family attached houses;
condominiums or cooperatives; mobile or manufactured homes; and any other type of
dwelling that is covered by an HO-01, HO-02, HO-03, HO-05, HO-06, HO-07 or HO-08
homeowners insurance policy.
The start date for the “Class Period” will be defined through discovery, which will reveal when

State Farm began using its automated claims processing methods and algorithms.

60. Upon information and belief, the Class contains tens of thousands of individuals.

State Farm is the largest issuer of homeowners insurance policies in North America, with

approximately 18% of market share, and is the biggest homeowners insurance provider in Illinois,

Indiana, Michigan, Missouri, and Ohio. It is the second biggest homeowners insurance provider

20
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 21 of 26 PageID #:21

in Wisconsin. The Midwest has close to one million Black homeowners. About 93% of

homeowners have home insurance, and approximately 5% of policyholders submit a claim on their

policy each year. The 2021 YouGov survey discussed supra by itself identified 134 Black State

Farm respondents who would be members of the Class under the proposed definition above. On

information and belief, the proposed Class is so numerous that joinder of all members would be

impracticable.

61. All members of the proposed Class have been subject to, and affected by, State

Farm’s claims processing methods. Questions of law and fact common to the proposed Class

predominate over questions affecting only individual members. These questions include, but are

not limited to, the following:

a. Whether State Farm uses automated methods of evaluating homeowners insurance

claims, and the nature, scope, and operation of those methods;

b. Whether State Farm’s automated methods of evaluating claims have caused racial
discrimination in violation of the FHA;

c. Whether there are statistically significant disparities between the processing time,
paperwork requests, and employee interactions for claims filed by Black and white
homeowners in the Midwest that are adverse to the Black homeowners;

d. Whether there is a legitimate business necessity for State Farm’s claims evaluation
methods;

e. Whether substantially equally or more valid alternative means of claim processing

are available that would eliminate or reduce the discriminatory impact; and

f. Whether an order for declaratory and injunctive relief is appropriate.

62. The claims of the individual named Plaintiff are typical of the claims of the

proposed Class and do not create any disabling conflicts with the interests of any other members

of the proposed Class.

21
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 22 of 26 PageID #:22

63. The individual named Plaintiff will fairly and adequately represent the interests of

the proposed Class. The Plaintiff and the proposed Class are represented by attorneys who are

qualified to pursue this litigation and have experience in class actions.

64. A class action is superior to other methods for the efficient and fair adjudication of

this controversy. A class action regarding the issues in this case does not create any problems of

manageability.

65. In the alternative, State Farm has acted or refused to act on grounds generally

applicable to the proposed Class, making appropriate final injunctive relief or corresponding

declaratory relief with respect to the proposed Class as a whole.

66. Finally, if the Court decides not to certify the Class under Rule 23(b)(2) or (3), class

certification is appropriate with respect to each of the common issues, including those identified

above and any other common issues that may be identified during the litigation, pursuant to Federal

Rule of Civil Procedure 23(c)(4).

V. CAUSES OF ACTION

COUNT ONE
VIOLATION OF THE FAIR HOUSING ACT, 42 U.S.C. § 3604(a) & (b)
(Individually and on Behalf of the Class)

67. Plaintiff, on behalf of herself and the Class Members, repeats and realleges every

allegation of the preceding paragraphs as if fully set forth herein.

68. Pursuant to the FHA, “[i]t is the policy of the United States to provide, within

constitutional limitations, for fair housing throughout the United States.” 42 U.S.C. § 3601.

69. Pursuant to 42 U.S.C. § 3604(a), “[i]t shall be unlawful . . . [to] otherwise make

unavailable or deny, a dwelling to any person because of race, color, religion, sex, familial status,

or national origin.”

22
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 23 of 26 PageID #:23

70. Pursuant to 42 U.S.C. § 3604(b), “[i]t shall be unlawful . . . [t]o discriminate against

any person in the terms, conditions, or privileges of sale or rental of a dwelling, or in the provision

of services or facilities in connection therewith, because of race, color, religion, sex, familial status,

or national origin.”

71. According to 24 C.F.R. § 100.70, promulgated by the U.S. Department of Housing

and Urban Development pursuant to 42 U.S.C. § 3604:

(b) [i]t shall be unlawful . . . to engage in any conduct relating to the provision of housing
or of services and facilities in connection therewith that otherwise makes unavailable or
denies dwellings to persons.

(d) Prohibited activities relating to dwellings under paragraph (b) of this section include,
but are not limited to:

...

(4) Refusing to provide . . . property or hazard insurance for dwellings or providing such .
. . insurance differently because of race . . . .

72. Homeowners insurance is a service rendered in connection with a sale of a

dwelling. State Farm provided homeowners insurance to the proposed Class representatives and

all proposed Class members, which they needed to maintain ownership of their home. State Farm’s

claims processing methods provide insurance coverage differently on the basis of race and have

thus resulted in discrimination with respect to the named Plaintiff as well as all proposed Class

members. State Farm’s claims processing methods violate Section 3604 of the FHA and constitute

actionable discrimination on the basis of race.

73. In addition to tangible financial harm due to State Farm’s delay in payments to

Black policyholders, the discriminatory effects of its claims processing methods have caused

Plaintiff and the proposed Class significant frustration, inconvenience, and humiliation on account

of their race.

23
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 24 of 26 PageID #:24

74. The proposed Class Representative and the proposed Class are aggrieved as defined

in the FHA by virtue of having been subject to State Farm’s discriminatory claims processing

methods, which artificially, arbitrarily, and unnecessarily subject the claims of Black policyholders

to greater scrutiny than those of White policyholders.

COUNT TWO
VIOLATION OF THE FAIR HOUSING ACT, 42 U.S.C. § 3605
(Individually and on Behalf of the Class)

75. Plaintiff, on behalf of herself and the Class Members, repeats and realleges every

allegation of the preceding paragraphs.

76. Pursuant to the FHA, “[i]t is the policy of the United States to provide, within

constitutional limitations, for fair housing throughout the United States.” 42 U.S.C. § 3601.

77. Pursuant to 42 U.S.C. § 3605, “[i]t shall be unlawful for any person or other entity

whose business includes engaging in residential real estate-related transactions to discriminate

against any person in making available such a transaction, or in the terms or conditions of such a

transactions, because of race . . . .”

78. “Residential real estate-related transaction” is defined to include “[t]he making or

purchasing of loans or providing other financial assistance . . . for . . . improving, repairing, or

maintaining a dwelling.” 42 U.S.C. § 3605(b)(1)(A).

79. Homeowners insurance companies operate to provide their policyholders the

financial assistance needed to improve, repair, and maintain their homes when property damage

and loss occurs. This financial assistance is facilitated by way of the insurance company’s claims

processing methods. The homeowners insurance claims process at State Farm thus constitutes the

sort of residential real estate-related transaction contemplated by Section 3605 of the FHA.

24
 Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 25 of 26 PageID #:25

80. As set forth above, State Farm uses claims processing methods that have a

discriminatory impact on Black policyholders, including on the named Plaintiff and all members

of the proposed Class. These claims processing methods artificially, arbitrarily, and unnecessarily

subject the claims of Black homeowners to greater scrutiny than those of white homeowners, and

thus result in greater delays and inconvenience to Black homeowners.

81. The named Plaintiff and the Class have been injured by State Farm’s discriminatory

conduct and have suffered damages as a result.

VI. JURY DEMAND

82. Plaintiff, on behalf of herself and the proposed class, hereby request a trial by jury.

VII. PRAYER FOR RELIEF

WHEREFORE, Plaintiff respectfully requests the following relief:

A. Certify this case as a class action and certify the named Plaintiff herein to be the
class representative, and counsel to be class counsel;

B. Enter a judgment pursuant to 42 U.S.C. § 3613 declaring the acts and practices of
State Farm complained of herein to be in violation of the FHA;

C. Grant a permanent or final injunction, pursuant to 42 U.S.C. § 3613 enjoining State

Farm and State Farm’s agents and employees, affiliates and subsidiaries, from
continuing to discriminate against Plaintiff and the members of the Class because
of their race through further use of their current claims processing methods;

D. Order State Farm, pursuant to 42 U.S.C. § 3613, to monitor and/or audit its claims
processing methods to ensure the cessation of discriminatory effects in claims
processing;

E. Order actual damages to the Plaintiff and the Class pursuant to 42 U.S.C. § 3613;

F. Award Plaintiff the costs of this action, including the fees and costs of experts,
together with reasonable attorneys’ fees, pursuant to 42 U.S.C. § 3613.

G. Grant the Plaintiff and the Class such other and further relief as this Court finds
necessary and proper.

25
Case: 1:22-cv-07014 Document #: 1 Filed: 12/14/22 Page 26 of 26 PageID #:26

Respectfully submitted,
/s/
Joshua Karsh
_________________________

Joshua Karsh
MEHRI & SKALET PLLC
1237 Judson Ave.
Evanston, IL 60202
Phone: (773) 505-7533
[email protected]
ARDC No. 6203096

Local Counsel for Plaintiff and the Proposed Class,

Pursuant to L.R. 83.15

FAIRMARK PARTNERS, LLP

Aisha Rich (Pro Hac Vice Forthcoming)
Alexander Rose (Pro Hac Vice Forthcoming)
Jamie Crooks (Pro Hac Vice Forthcoming)
1825 7th St NW, #821
Washington, DC 20001
Phone: (301) 458-0564
[email protected]
[email protected]
[email protected]

CENTER ON RACE, INEQUALITY, AND THE

LAW AT NEW YORK UNIVERSITY SCHOOL
OF LAW
Deborah N. Archer (Pro Hac Vice Forthcoming)
Jason D. Williamson (Pro Hac Vice Forthcoming)
139 MacDougal Street
New York, NY 10012
Phone: (212) 998-6882
[email protected]
[email protected]

Counsel for Plaintiff and the Proposed Class

26