Generative AI

2023 Report

CREATED BY:
Generative AI 2023 Report

Contents

3 Foreword 54 Section 6.
Trust and quality
4 Introduction
57 Section 7.
9 Section 1. Non-use of generative AI tools
Which tools have you heard of?
61 Section 8.
18 Section 2. Generative AI future insights
Why did you decide to start using generative AI?
67 Conclusion
30 Section 3.
What do you mainly use generative AI tools for? 68 Contributors

45 Section 4.
Main generative AI tool of choice, benefits, and challenges

52 Section 5.
How much has your generative AI tool of choice impacted
your productivity and efficiency?

2
Generative AI 2023 Report

Foreword
It’s with great pleasure that I introduce the inaugural Generative The subjects and themes explored in this report include techniques,
AI report. Since establishing the AI Accelerator Institute (AIAI), I’ve trends, motivators, and challenges. This work represents the
been immensely proud of the contributions AIAI has made collective insights of experts in the field, whose contributions we
in providing expert insights to members of the artificial greatly appreciate. Their dedication, time, and knowledge have
intelligence community. been invaluable to AIAI, and this report wouldn’t have been possible
without them.
Generative AI is a branch of artificial intelligence and machine
learning that primarily focuses on generating new data from The aim of this report is to comprehend the primary motivators
existing information. Its applications are versatile and cross-cutting, molding the industry, identify the significant limitations, and
impacting sectors such as entertainment, healthcare, retail, design, spotlight upcoming trends that companies should be aware of to
and more. maintain their competitive edge.

Numerous factors have contributed to the rise of generative AI in AIAI anticipates this report will enhance our shared
recent years. The advancements in deep learning technologies understanding of the generative AI market in 2023 and lay the
have been a key driver, with algorithms now capable of generating groundwork for future collaborations with companies, academics,
an impressive array of realistic and creative outputs, from images and professionals globally.
to music, and more.

In addition, the evolution of hardware capabilities has significantly

influenced the growth of generative AI, enabling the processing of
Jordan Dunne
vast amounts of data quickly and efficiently. The expansion of data
Founder, AI Accelerator Institute
sets, including an array of open-source options, has provided the
necessary training grounds for these algorithms to improve their
accuracy and performance.

3
Introduction

4
 Generative AI 2023 Report

Introduction
Why did we create this report? How we collected the data
We developed the Generative AI report to uncover insights on To collect the quantitative data to build this report, we surveyed
the primary forces fueling the rising adoption of generative AI, hundreds of generative AI users across the AIAI community and
understand the most pressing challenges in integrating generative beyond. This gave us an understanding of how the market is currently
AI at present, and identify the emerging trends on the horizon. viewing the technology. We also wanted to know what their predictions
are for the future.
Remaining at the forefront of generative AI innovations could be the
differentiation that sets you apart from your competitors. In this report, we offer a qualitative overview of the respondents’
answers to offer insight into the answers.
In this report, we delve into the generative AI applications that
businesses are emphasizing and introducing, and we explore the The findings point towards a significant portion of our participants
dynamics prompting shifts in the marketplace. residing in North America and Europe, a predictable outcome given the
concentration of industry headquarters in these areas. However, we
also see contributions from diverse global regions, like South America
Europe and Africa, which is vital for a balanced representation of data.
North America

31.3%
33%
24.1% Asia

7.1%

2.7% 1.8%

South
Hint: America

If you’re pressed for Africa Australasia

time, you can find the
TL;DR on page 68 5
Generative AI 2023 Report

By industry
We asked our respondents which industry they operated in, and
these were the results:

Those who answered “other” added their industries:

Computer hardware - 2.7%
1. Aerospace
2. Food industry
Computer software - 33.6%
3. Clean energy storage
4. Marketing and Communications
Cybersecurity - 3.6%
5. MR
6. Public Service
Education - 11.8%
7. Advertising
8. AI/ML
Healthcare - 10.1%
9. Business Consulting
10. IT Staffing Services
Telecommunications - 1.8%
11. Compute Architecture
12. Public Housing
Retail - 1.8%
13. Data Management & Annotation
14. Design
Transportation - 3.6% 15. Consulting
16. Legal/Law
Manufacturing - 5.5% 17. UX Design & Research
18. Research & Development
Banking/Finance - 6.4%

Other - 15.5%

6
Generative AI 2023 Report

By main business function By number of employees

We asked our respondents what their main business function was, As you can see, the majority of the companies that took part in
and these were the results: this report have between 1-10 employees (30.6%), followed by
companies with over 500 employees (21.6%). It’s also interesting to
note that we have a somewhat even spread of employee numbers
Management
Other per company.
17.3%
20.9%
Having information from companies with such different numbers
of employees gives us insights into what companies of all sizes
prioritize when using generative AI tools.

Marketing 500+
12.7% 21.6% 1-10
Business function 30.6%

Design
14.5%
Sales
9.1% Number of
Employees

Research and development 251-500

25.5% 13.5%
11-20
12.5%

51-100
12.5% 21-50
9.3%

7
Generative AI 2023 Report

By company revenue By stage

<$50 million 61.9% 25.3%

Early pre-product market fit

$51-$100 million 13.6%

Early post-product market fit 13.6%

$101-$200 million 4.5%

Mid-growth with an established
19.9%
go-to-market team
$201-$300 million 2.7%

Late growth/scale-up 17.8%

$301-$400 million 0.9%

>$500 million Enterprise 23.4%

16.4%

When it comes to company revenue, you can see that the majority As the results show, the highest percentage of respondents’
of respondents were located in the band with less than $50 million companies are at an early pre-product market fit level, having
(61.9%), with the next most common band being companies with a 25.3% of votes. Following closely behind are companies at an
revenue of over $500 million (16.4%). enterprise level (23.4%) and mid-growth with an established
go-to-market team (19.9%).
It is, however, great to see a diverse representation of company
revenue, which points towards the usefulness of generative AI being
beyond a potential cost saving feature.

8
Section 1

Which tools have you heard of?

9
Computer Vision in Healthcare

Generative artificial intelligence (AI) lets users quickly create

new content based on a wide variety of inputs. These can be text,
images, animation, sounds, 3D models, and more.
Generative High
These systems use neural networks to identify patterns in existing Denoising
Adversarial Quality
Diffusion Models
data, producing fresh and unique content. One significant Networks Samples
advancement in generative AI is the capacity to utilize various
learning methods, like unsupervised or semi-supervised learning,
during training.

This allows individuals to efficiently use vast amounts of unlabeled

data to construct foundation models. These models serve as the Mode
Fast
groundwork for multifunctional AI systems. Coverage
Sampling
Diversity

1. Quality

Mainly important for applications that interact with users directly,

a high-quality generation output is vital. In speech generation,
for example, having poor speech quality means it’ll be difficult to Variational Autoencoders,
Normalizing Flows
understand, and in image generation, outputs need to be visually
indistinguishable from natural images.

2. Diversity

Good generative AI models can capture minority modes in their

data distribution without compromising on quality. This leads to a
minimization of undesired biases in learned models.

3. Speed

A wide variety of interactive applications need fast generation, like

real-time image editing for content creation workflows.

10
Generative AI 2023 Report

How do you develop generative

AI models?
There are several types of generative models; combining their positive attributes will lead to even more powerful models:

Diffusion models Variational autoencoders (VAE) models

Also known as denoising diffusion probabilistic models (DDPMs), Consisting of two neural networks: the encoder and the decoder. When
these determine vectors in latent space through a two-step process VAE models are given an input, the encoder converts it into a smaller
when in training. and denser representation of the data.

1. Forward diffusion. The compressed representation of data keeps the information

This process slowly adds random noise to training data. needed for a decoder to then reconstruct the original input data
while discarding anything irrelevant. Both encoder and decoder work
2. Reverse diffusion.
together to learn a simple and efficient latent data representation,
This process reverses the noise and reconstructs data samples.
allowing users to sample new latent representations that can be
New data is created by running the reverse denoising process from mapped through the decoder to create new data.
entirely random noise.
VAE models can create outputs, like images for example, faster but
Diffusion models can, however, take longer to train than variational they won’t be as detailed as the ones from diffusion models.
autoencoder (VAE) models. But the two-step process allows for
hundreds, and even an infinite number, of layers to be trained,
meaning diffusion models tend to offer the highest quality of output
when you’re building generative AI models.

Also categorized as foundation models, diffusion models are large-

scale, they’re flexible, and tend to be the best for generalized use
cases. Their reverse sampling process does, however, make running
them a slow and lengthy process.

11
Generative AI 2023 Report

Generative adversarial network (GAN) models

Before diffusion models, GANs were the most commonly used methodology. These models place two neural networks
against each other.

1. Generator.
Creates new examples.

2. Discriminator.
Learns to separate created content as real or fake.

GANs can offer high-quality samples and they often create outputs quickly; the sample diversity, however, is weak, and
GANs are better suited for domain-specific data generation.

In the report, we started with a simple question: Do you use generative AI tools?

Yes
88.2%

No
11.8%

88.2% of respondents replied with yes, and we wanted to investigate which generative AI tools they previously heard of.

In this section, we’ve highlighted the top three answers we received:

12
Generative AI 2023 Report

ChatGPT
Developed by OpenAI, ChatGPT allows users to have free access to basic artificial intelligence content generation. Its premium
subscription, ChatGPT Plus, is marketed to users who need extra processing power and want early access to new features.

Key features Pros

• Language fluency • A free version for the general public

• Personalized interactions • Offers more accurate answers and natural

• Conversational context interactions

• Language translation • The API lets developers embed a ChatGPT

functionality into apps and products
• Natural language understanding
• Completion and suggestion of text
• Open-domain conversations

Cons
• Can’t access data after September 2021, but
plugins may help fix this issue
Use case
• Can be prone to errors and misuse
• Chatbot
• Content generation

Pricing
• A free version is available
• Paid membership: begins at $0.002 per 1,000
prompt tokens

13
Generative AI 2023 Report

GPT-4
It creates human-like text responses to both word prompts and questions. Each response is unique, allowing you to enter the same
query as many times as you want and get different responses every time.

The latest version of this large language model, GPT-4, has been marketed as more accurate and inventive than its previous
iterations while being safer and more stable.

Key features Cons

• Multilingual ability • Can have biases
• Human-level performance • Can provide wrong answers
• 100 trillion parameters • Image inputs unavailable for public use
• Enhance steerability
• Image input ability
• Factual performance improved

Pricing
• A free version is available
• $0.03 per 1,000 prompt tokens
Use case
• $0.06 per 1,000 completion tokens
• Large language model
• Paid membership: $20/month

Pros
• A cost-effective solution
• Consistent and reliable time saver
• GPT-4 has more extensive safety checks and training
than previous versions

14
Generative AI 2023 Report

What’s the difference between GPT and ChatGPT?

ChatGPT is the app and GPT is the brain behind it.

Simply put, this is the difference between GPT and ChatGPT.

GPT ChatGPT
Large language Chatbot app
model (LLM) powered by GPT

• Gived AI apps the ability • Relies on a set of

to generate text
VS GPT’s parameters

• One of the largest • Optimized for

neural networks, with dialogue
hundreds of billions of
• Has content filters
parameters

For efficiency purposes, in this report, we use ChatGPT as a blanket term for OpenAI’s offerings.

15
Generative AI 2023 Report

Bard
Both a content generation tool and a chatbot, Bard was developed by Google. It uses LaMDA, which is a transformer-based model,
and it’s often seen as ChatGPT’s counterpart.

By May 10, Google opened up access to Bard for everyone and added functionalities such as image processing, coding features, and
app integration. This enabled a broad spectrum of users, including developers and marketers from around the globe, to leverage
Bard for their professional tasks.

Unlike ChatGPT, which has an information cutoff in September 2021, Google Bard has live internet connectivity, allowing it to provide
real-time information. According to Sundar Pichai, CEO of Google and Alphabet, Bard strives to merge the expansive knowledge of
the world with the capabilities of large language models, generating high-quality responses by sourcing information from the web.

Notably, Google currently views Bard as an ongoing experiment.

Key features Pros

• Rating system for user responses • Pre-tested extensively
• Can help with tasks related to software • A transparent and ethical approach to AI development
development and programming
• Built on LaMDA
• Available through individual Google accounts
Cons
• Only available in English
• Not available through Google accounts managed
Use case by a Google Workspace admin
• Chatbot • No conversational history
• Content generation

Pricing
• Free

16
Generative AI 2023 Report

Midjourney
Midjourney stands as a cutting-edge AI art interface, tapping into generative algorithms to fuel artistic creativity. It helps artists to
create distinct and captivating pieces, capitalizing on advanced machine learning methodologies.

Offering both art prompts and ideas, Midjourney can even mold full-fledged artworks in response to user preferences. Its intricate
neural network has been shaped by comprehensively studying a variety of artistic datasets, paintings, sketches, and photos.

Midjourney appeals to a diverse audience, from seasoned artists who want a fresh point of view to novices wanting to get started.

Key features Cons

• High-resolution images • AI still under development

• Great image composition • High usage costs

• Collaborative potential • Platform not as user-friendly as other options

• Professional applications of images

Pricing
Use case • Basic: $10 per month, 3.3 fast hours
• Endless prompt generation • Standard: $30 per month, 15 fast hours per month
• Offers big style diversity • Pro: $60 per month, 30 fast hours
• Efficient iteration • Mega: $120 per month, 60 fast hours

Pros
• Endless prompt generation
• Offers big style diversity
• Efficient iteration

17
Section 2

Why did you decide to

start using generative AI?

18
Generative AI 2023 Report

As the large majority of respondents (88.2%) use generative AI tools, we wanted to understand why they decided to
start using them.

To reduce costs
Other 2.7%
15.5%

Out of
curiosity
22.7%

To speed up
To write
work
software
48.2%
5.5%

To increase
personalization To create content for
0.9% daily work
4.5%

19
Generative AI 2023 Report

To speed up work
The plurality of respondents, 48.2%, said they started using generative AI to speed up their work.

Thanks to its capability of producing and assisting in decision-making across several areas, generative AI can considerably speed
up work processes in companies. It enhances human input and makes sure that time-consuming tasks are completed in a fraction
of the time it typically takes.

With artificial intelligence technologies progressively being integrated into workplaces, we can reasonably expect that businesses will
operate at an even quicker pace, which will make adaptability and speed essential for success.

Let’s take a look at ways in which generative AI can help speed up work processes:

Generative AI can quickly analyze large amounts of data to produce insights or suggestions. In finance, for
Improving
1. decision-making
example, AI can generate investment strategies by considering thousands of data points and trends much
quicker than a human analyst could. This leads to faster and potentially more accurate decisions.

When it comes to architecture or product design, generative AI can produce multiple design variations in
Enhancing creativity
2. and design
minutes according to project needs. This means designers can quickly iterate and refine ideas, cutting down
the time traditionally required in the design phase.

Generative AI can draft articles, generate graphics, or produce video content at an impressive speed. This
Streamlining content
3. creation
quick content-generation ability can be particularly useful for industries like journalism, advertising, and
entertainment.

AI chatbots can offer real-time answers to customer queries, which greatly reduces or even eliminates
Providing instant
4. answers to customers
wait times. Whether it’s helping with troubleshooting, product information, or general inquiries, immediate
feedback enhances customer experience.

Speeding up research In sectors like biotechnology, for example, AI can predict molecule interactions or simulate experiments at a
5. and development much quicker rate than traditional methods. This means reduced time-to-market for new drugs or materials.

20
Generative AI 2023 Report

Tasks like data entry, scheduling, and basic administrative duties can be completed faster and more
Increasing task
6. automation efficiency
efficiently using generative AI. When these repetitive tasks are addressed quickly, businesses can focus on
more complex and strategic endeavors.

Completing real-time Generative AI can rapidly predict market trends, customer preferences, or inventory needs. This instant
7. forecasting forecasting helps businesses to make swift decisions, adjust marketing strategies, or manage stock.

AI-based training programs can be generated based on individual needs, which makes sure that employees
Generating training
8. modules
are brought up to speed faster. Through this tailored content, training durations are minimized, and
efficiency is boosted.

Generative AI can quickly screen candidate profiles, matching skills and qualifications with job requirements.
Speeding up
9. recruitment processes
This speeds up the shortlisting process and helps companies hire employees faster, which reduces vacant
position downtimes.

Enhancing AI systems can detect and neutralize threats in real-time, making sure that business operations are
10. cybersecurity uninterrupted. A fast response to potential threats leads to less downtime and swift work processes.

21
Generative AI 2023 Report

Out of curiosity
22.7% of respondents stated they started using generative AI technologies out of curiosity.

Using these tools to satisfy curiosity can lead to unexpected artistic expressions, discoveries, or even fresh perspectives. Generative
AI can be much more than just utilitarian; when human curiosity meets cutting-edge technology, the possibilities are endless.

There are a few ways that generative AI technologies can be used out of curiosity and artists, for example, can use this technology
to produce unique pieces of art, music, or literature. Simple and basic inputs can generate astonishing variations of artworks, songs,
or stories, offering unexpected and imaginative creations.

Generative AI can suggest plots, themes, or design concepts to help writers or designers. Through these suggestions, creative
individuals can have new ideas or overcome creative blocks.

By experimenting with generative AI tools, you can refine your coding, problem-solving, and analytical skills. This hands-on
approach can help in understanding the intricacies of artificial intelligence and its other potential applications.

Sometimes, however, it can be fun to simply have an open-ended exploration without a specific goal. This can lead to unexpected
innovations or findings.

22
Generative AI 2023 Report

To write software
Generative AI has massive potential in software development, and 5.5% of respondents are doing just that.

Generative AI’s role in software development is paving the way for faster, more efficient, and more intuitive software creation
processes. This technology can significantly improve writing, testing, and optimizing software, leading to a transformation in how
software is conceptualized, developed, and deployed.

Let’s have a look at how generative AI is changing software development:

This technology can help developers by auto-generating bits of code based on context. By understanding
Generating and
1. auto-completing code
the objective and the existing code structure, AI can suggest or even write snippets, which speeds up the
development process.

By analyzing big code repositories, generative AI models can easily predict where bugs could happen –
2. Detecting bugs and even suggest potential fixes. This proactive approach can lead to more stable software and reduce
debugging time.

AI can simulate a variety of user behaviors and scenarios to help test software. This makes sure that
3. Testing software comprehensive testing is completed in a fraction of the time, which provides strong and reliable software
applications.

Generative AI can help with the integration of many APIs by understanding their documentation and
Providing API
4. integrations
generating appropriate integration code, simplifying the process of adding new functionalities to
applications.

Enhancing user Generative design tools can create multiple UI variations based on given parameters. Developers and
5. interface (UI) design designers can streamline the UI creation process by choosing or iterating from these designs.

Providing personalized Generative AI tools can analyze user behavior and feedback, suggesting or even implementing UX
6. user experience (UX) improvements so the software can then be adapted to meet individual user needs and preferences.

Managing and Artificial intelligence can help with structuring, querying, and optimizing databases. When predicting
7. optimizing databases potential bottlenecks or inefficiencies, AI can ensure straightforward and efficient data operations.

Generative AI can simulate cyber-attacks or probe software for vulnerabilities. This helps developers
8. Improving security
strengthen their applications, as they can understand and predict potential security flaws.

23
Generative AI 2023 Report

To create content for daily work

4.5% of respondents started using generative AI to create content for their daily work.

These technologies are reshaping the daily work process content landscape, as they provide quick, tailored, and efficient content
generation. This lets professionals focus on creative or strategic aspects of their tasks.

As artificial intelligence keeps evolving, its integration into everyday work tasks is likely to become even more prevalent, simplifying
the content generation process and enhancing overall productivity.

Generative AI can quickly draft reports, summaries, or other documents based on provided data or
Drafting reports and
1. documents
guidelines. Because you don’t start from scratch and have a foundational draft, you can refine it as
needed and streamline your work.

12.7% of respondents’ main business function is marketing – and generative AI can greatly help in analyzing
Content personalization
2. for marketing
user preferences and behavior. It can tailor content to individual users by creating personalized email
campaigns or customized product recommendations on e-commerce platforms.

For news outlets and publishers, artificial intelligence can draft news articles or updates, especially for
3. Automated journalism repetitive content like sports scores or financial updates. This lets human journalists focus on in-depth
analyses and features.

Generative AI tools can generate a variety of visual content, from website banners to product mock-ups. For
4. Graphic design
daily tasks, like social media posts, AI can deliver many design options, easing the rapid content roll-out.

AI can process large amounts of literature or data to generate summaries or insights in academia. Instead
5. Research summaries of filtering through numerous papers, professionals can receive a condensed overview, which accelerates
the research process.

24
Generative AI 2023 Report

Drafting emails, proposals, or other communications is much faster with generative AI. The technology
6. Email writing
uses key points or themes to give users a well-structured draft, streamlining daily communication tasks.

11.8% of respondents work in the education sector. For trainers, educators, or e-learning platforms, AI can
7. Educational content
generate quizzes, assignments, or study summaries based on provided course material.

For content-based websites, generative AI can create article drafts, topic suggestions, or even SEO-
8. Article creation
optimized content. This can be especially useful for maintaining daily content schedules.

Social media managers can use artificial intelligence to create post captions, responses to comments, or
Social media
9. management
content suggestions based on trends. This means you can have consistent engagement without needing
continuous manual input.

AI tools can process recordings or notes to create succinct minutes or action points. This reduces
Meeting notes
10. and minutes administrative load after meetings and helps participants have a clear understanding of what was
discussed.

25
Generative AI 2023 Report

To reduce costs
Through using generative AI, businesses can have a competitive advantage by innovating and saving on costs.

With automating, optimizing, and predicting, companies can easily streamline operations, reduce waste, and make sure they get the
best value for their outgoings. AI technology keeps evolving, meaning that its potential for cost savings will only grow.

Here are a few ways that AI can help companies save on costs:

Generative AI can create many design alternatives by defining specific constraints and parameters.
Product design and
1. prototyping
Designers can use AI to rapidly generate hundreds of options in seconds instead of days or even weeks,
which reduces both time and material costs.

Generating content, such as advertising, web designs, or articles, can be a resource-intensive process.
Generative AI models can generate human-like text, images, or even videos. The automation of part of
2. Content creation
the content creation process helps businesses drastically reduce the costs associated with hiring multiple
content creators, graphic designers, and videographers.

Personalization and Generative AI tools can create personalized content for users based on their preferences and behavior. This
3. customer engagement personalization improves user engagement and can result in higher conversion rates.

A variety of businesses face the challenge of repetitive and mundane tasks, like data entry, report
Repetitive task
4. automation
generation, and simple customer service inquiries. Generative AI can automate these processes, leading to
significant savings in labor costs and increasing overall employee efficiency.

Generative AI models can help with drug discovery, materials science, and other sectors with intensive
Enhanced research and research and development. By predicting molecular structures, testing potential scenarios, or simulating
5. development experiments, AI can severely minimize the number of physical tests required, which accelerates timelines
and saves on costs.

26
Generative AI 2023 Report

Generative AI-powered chatbots can handle a wide range of customer inquiries without employee
Customer service and intervention. These systems can offer instant answers at any time of day, which leads to improved
6. support customer satisfaction while drastically reducing the need for large customer service teams working
around the clock.

Generative AI can be used to simulate different business scenarios, which helps companies to make better-
informed decisions about inventory management, sales strategies, and more. By accurately predicting
7. Improved forecasting
demand or potential business disruptions, companies can reduce waste, avoid overstocking, and optimize
supply chains.

By using Generative AI to create personalized learning paths for employees, businesses don’t need to invest
8. Training and education heavily in training programs, seminars, or courses. These AI-driven platforms can adapt to each individual’s
learning pace and needs, reducing the time and cost of training.

Screening candidates, processing applications, and performing initial interviews can be time-consuming
Recruitment and
9. human resources
and expensive. Generative AI tools can analyze large amounts of applications, predict the fit between
candidates and jobs, and even automate the initial communication between companies and applicants.

Generative AI can simulate cyberattacks and help companies identify vulnerabilities in their systems.
Enhancing This proactive approach can prevent expensive breaches and make sure there aren’t any interruptions in
10. cybersecurity business continuity. AI-driven systems can also monitor networks in real time, identifying and countering
threats faster than human-only teams.

27
Generative AI 2023 Report

To increase personalization
The increasing integration of generative AI into personalization is changing how businesses and platforms interact with and serve
their users. By offering highly tailored experiences, products, and services, companies can enhance user satisfaction and encourage
deeper loyalty and trust.

Here’s how this technology enhances personalization:

Generative AI can tailor the shopping experience by analyzing user behavior, preferences, and purchase
1. E-commerce experience history. It can also recommend products, offer personalized discounts, or even generate custom product
designs, making online shopping a better experience according to individual preferences.

Streaming platforms and social media platforms, for example, can use generative AI to curate content feeds.
Content
2. recommendations
By understanding user preferences, these platforms can offer highly relevant content, such as articles or
posts to improve user engagement.

Students can have a more personalized learning path with generative AI. The technology can assess
3. Learning and education students’ strengths, weaknesses, and learning paces, offering tailored lessons, assignments, or resources for
optimal learning outcomes.

Companies can use generative AI to create personalized marketing messages, email campaigns, or
Marketing and
4. advertising
advertisements. Understanding individual user demographics, interests, and behaviors, helps to make
marketing more effective.

Generative AI can create custom workout plans, diet charts, or even mental health exercises by analyzing
5. Health and fitness a person’s health data, goals, and preferences. This leads to a more effective and sustainable wellness
journey.

28
Generative AI 2023 Report

Chatbots and support systems powered by generative AI can offer personalized solutions based on a
6. Customer support
user’s past interactions, purchase history, and preferences, for faster and better issue resolution.

Companies can use generative AI to analyze customer feedback, reviews, and preferences to design
7. Product development
products or services. Products can then meet market demand and resonate with target audiences.

Banks and financial institutions can utilize generative AI to offer personalized financial advice, investment
8. Financial services
strategies, or loan options based on individual financial behavior, needs, and goals.

Generative AI can create personalized event agendas, travel itineraries, or experiences. It can help plan a city
9. Event planning
tour based on interests or other more personalized ideas according to every individual user.

User interface and Generative AI can adapt and redesign software or website interfaces based on user behavior. This offers
10. experience (UI/UX) users a smoother, more intuitive, and more engaging digital experience.

29
Section 3

What do you mainly use

generative AI tools for?

30
Generative AI 2023 Report

We asked respondents the main use of generative AI of generative AI tools, and the majority highlighted text applications
(40.8%). In second place we see conversational applications (29.6%) and in third place data augmentation (13.3%).

40.8%

29.6%

13.3%

7.1% 7.1%

2.1%

Visual/video Audio Text Conversational Data Other

applications applications applications applications augmentation

Audio applications
Generative AI audio models use machine learning techniques, artificial intelligence, and algorithms to create new sounds from
existing data. This data can include musical scores, environmental sounds, audio recordings, or speech-to-sound effects.

After the models are trained, they can create new audio that’s original and unique. Each model uses different types of prompts to
generate audio content, which can be:
• Environmental data
• MIDI data
• User input in real-time
• Text prompts
• Existing audio recordings

31
Generative AI 2023 Report

There are several applications of generative AI audio models:

1. Data sonification 5. Sound effects creation and synthesis

Models can convert complex data patterns into auditory Models can enable the synthesis of unique and realistic sounds,
representations, which lets analysts and researchers understand like instruments, abstract soundscapes, and environmental effects.
and explore data through sound. This can be applied to scientific They can create sounds that copy real-world audio or completely
research, data visualization, and exploratory data analysis. new audio experiences.

2. Interactive audio experiences 6. Audio captioning and transcription

Creating interactive and dynamic audio experiences, models can Helping to automate speech-to-text transcription and audio
generate adaptive soundtracks for virtual reality environments captioning, models can greatly improve accessibility in several
and video games. The models can also respond to environmental media formats like podcasts, videos, and even live events.
changes or user inputs to improve engagement and immersion.

7. Speech synthesis and voice cloning

3. Music generation and composition
You can clone someone’s voice through generative AI models and
Creating musical accompaniment or composing original music create speech that sounds exactly like them. This can be useful for
pieces is easy for these models; they can learn styles and patterns audiobook narration, voice assistants, and voice-over production.
from existing compositions to generate rhythms, melodies, and
harmonies.
8. Personalized audio content

4. Audio enhancement and restoration Through the use of generative AI models, you can create
personalized audio content tailored to individual preferences. This
You can restore and enhance audio recordings with generative AI, can range from ambient soundscapes to personalized playlists or
which lets you reduce noise, improve the overall quality of sound, even AI-generated podcasts.
and remove artifacts. This is useful in audio restoration for archival
purposes.

32
Generative AI 2023 Report

How do generative AI audio models work? • Waveform sampling. WaveNet starts with an input waveform,
usually a sequence of audio samples, processed through
Like other AI systems, generative audio models train on vast data
multiple convolutional layers.
sets to generate fresh audio outputs. The specific training method
can differ based on the architecture of each model. • Dilated convolution. To recognize long-spanning dependencies
in audio waveforms, WaveNet employs dilated convolutional
Let’s take a look at how this is generally done by exploring two
layers. The dilation magnitude sets the receptive field’s size in
distinct models: WaveNet and GANs.
the convolutional layer, helping the model distinguish extended

WaveNet patterns.

Created by Google DeepMind, WaveNet is a generative audio model • Autoregressive model. Functioning autoregressively, WaveNet

grounded on deep neural networks. Using dilated convolutions, it sequentially generates audio samples, each influenced by its

creates great-quality audio by referencing previous audio samples. predecessors. It then forecasts the likelihood of the upcoming

It can produce lifelike speech and music, finding applications in sample based on prior ones.

speech synthesis, audio enhancement, and audio style adaptation. • Sampling mechanism. To draw audio samples from the model’s
Its operational flow consists of: predicted probability distribution, WaveNet adopts a softmax
sampling approach, ensuring varied and realistic audio output.

• Training protocol. The model undergoes training using a

Residual Layer N
maximum possibility estimation technique, which is designed
Residual to increase the training data’s probability when it comes to the
Unit
model’s parameters.
M feature layers

1x1 1x1
ReLU ReLU SoftMax Output probability
Residual Convolution Convolution distribution
outputs

Residual Layer 1

Residual
Unit
M feature layers

Casual
Convolution

Audio Input
33
Generative AI 2023 Report

Generative Adversarial Networks (GANs) • Adversarial loss. GANs aim to reduce the adversarial loss,
which is the gap between real audio sample distributions
A GAN encompasses two neural networks: a generator for creating
and fake ones. This minimization rotates between the
audio samples and a discriminator for judging their authenticity.
generator’s enhancements for more authentic output and
Here’s an overview:
the discriminator’s improvements in differentiating real from
• Architecture. GANs are structured with a generator and generated audio.
discriminator. The former ingests a random noise vector,
• Audio applications. GANs have various audio purposes, such as
outputting an audio sample, while the latter evaluates the
music creation, audio style modulation, and audio rectification.
audio’s authenticity.
For music creation, the generator refines itself to form new
• Training dynamics. The generator creates audio samples from musical outputs. For style modulation, it adapts the style from
random noise during training and the discriminator’s task is one sample to another. For rectification, it’s trained to eliminate
to categorize them. Working together, the generator refines noise or imperfections.maximum possibility estimation
its output to appear genuine to the discriminator, and this technique, which is designed to increase the training data’s
synchronization is executed by reducing the binary cross- probability when it comes to the model’s parameters.
entropy loss between the discriminator’s findings and the actual
labels of each sample.

Discriminator Model
Random Input
Vector Real Examples
(Random Noise)

Generator Model

Generated Fake Binary

Example Classification

Real Fake
Update Model

34
Generative AI 2023 Report

Text applications
Artificial intelligence text generators use AI to create written copy, 3. Summarization
which can be helpful for applications like website content creation,
report and article generation, social media post creation, and more. Helpful for text summarization, models provide concise and easy-
to-read versions of information by highlighting the most important
By using existing data, these artificial intelligence text generators
points. This is useful when it comes to summarizing research
can make sure that content fits tailored interests. They also help
papers, books, blog posts, and other long-form content.
with providing recommendations on what someone will most be
interested in, from products to information.
4. Chatbot and virtual assistants
There are several applications of generative AI text models:
Both virtual assistants and chatbots use text generation models
1. Language translation to be able to interact with users in a conversational way. These
assistants can understand user queries and offer relevant answers,
These models can be used to improve language translation alongside providing personalized information and assistance.
services, as they can analyze large volumes of text and generate
accurate translations in real time. This helps to enhance SEO-optimized content
5.
communication across different languages.
Text generators can help to optimize text for search engines. They
2. Content creation can decide on the meta description, headline, and even keywords.
You can easily find out the most search topics and their keyword
Perhaps one of the most popular applications, content creation volumes to make sure you have the best-ranking URLs.
refers to blog posts, social media posts, product descriptions,
and more. Models are trained on large amounts of data and can
produce high-quality content very quickly.

35
Generative AI 2023 Report

How do generative AI text models work? Transformers emerged as a solution to this dilemma. With the
increase in data volume and architectural complexity, transformers
AI-driven content generators use natural language processing
provide advantages like parallel processing capabilities. They’re
(NLP) and natural language generation (NLG) techniques to create
experienced at recognizing long patterns, which leads to stronger
text. These tools offer the advantage of improving enterprise
and more nuanced language models.
data, tailoring content based on user interactions, and crafting
individualized product descriptions. Simplified, the steps to text generation look like this:

• Data collection and pre-processing. Text data gathering,

Algorithmic structure and training
cleaning, and tokenization into smaller units for model inputs.
Content-based on NLG is crafted and structured by algorithms.
• Model training. The model is trained on token sequences, and
These are typically text-generation algorithms that undergo
it adjusts its parameters in order to predict the next token in a
an initial phase of unsupervised learning. During this phase, a
sequence according to the previous ones.
language transformer model immerses itself in vast datasets,
extracting a variety of insights. • Generation. After the model is trained, it can create new text
by predicting one token at a time based on the provided seed
By training on extensive data, the model becomes skilled in
sequence and on tokens that were previously generated.
creating precise vector representations. This helps in predicting
words, phrases, and larger textual blocks with heightened context • Decoding strategies. You can use different strategies, such
awareness. as beam search, op-k/top-p sampling, or greedy coding to
choose the next token.
Evolution from RNNs to transformers
• Fine-tuning. The pre-trained models are regularly adjusted on
While Recurrent Neural Networks (RNNs) have been a traditional particular tasks or domains to improve performance.
choice for deep learning, they often have difficulty in modeling
extended contexts. This shortcoming comes from the vanishing
gradient problem.

This issue happens when deep networks, either feed-forward or

recurrent, find it difficult to relay information from the output layers
back to the initial layers. This leads to multi-layered models either
failing to train efficiently on specific datasets or settling prematurely
for less-than-ideal solutions.

36
Generative AI 2023 Report

Conversational applications
Conversational AI focuses on helping the natural language 3. Natural language generation (NLG)
conversations between humans and AI systems. Leveraging
technology like NLG and Natural Language Understanding (NLU), it To generate human-like answers in real time, conversational AI
allows for seamless interactions. systems use NLG techniques. By taking advantage of pre-defined
templates, neural networks, or machine learning models, the
There are several applications of generative AI conversational
systems can create meaningful and contextually appropriate
models:
answers to queries.

1. Natural Language Understanding (NLU)

4. Dialogue management

Conversational AI uses sophisticated NLU techniques to understand

Using strong dialogue management algorithms, conversational AI
and interpret the meanings behind user statements and queries.
systems can maintain a context-aware and coherent conversation.
Through analyzing intent, context, and entities in user inputs,
The algorithms allow AI systems to understand and answer user
conversational AI can then extract important information to
inputs in a natural and human-like way.
generate appropriate answers.

2. Speech recognition

Conversational AI systems use advanced algorithms to transform

spoken language into text. This lets the systems understand and
process user inputs in the form of voice or speech commands.

37
Generative AI 2023 Report

How do generative AI conversational models work?

Backed by underlying deep neural networks and machine learning, a typical conversational AI flow involves:

• An interface that lets users input text into the system or automatic speech recognition, which is a user interface that
transforms speech into text.

• Natural language processing extracts users’ intent from text or audio input, translating text into structured data.

• Natural language understanding processes data based on context, grammar, and meaning to better understand entity and
intent. It also helps it to act as a dialogue management unit in order to build appropriate answers.

• An AI model predicts the best answer for users according to the intent and the models’ training data. Natural language
generation infers from the processes above to form an appropriate answer to interact with humans.

Automated Speech
Recognition Sure! I can give you
Hi, I’d like to know what a description of what
generative AI means. generative AI means.
LISTENING
Dialog Natural Language
Management Generation

FORMING OFFERING
RESPONSE RESPONSE
Natural Language
Understanding

COMPREHENDING

38
Generative AI 2023 Report

Data augmentation Variational Autoencoders (VAEs)

Through using artificial intelligence algorithms, especially Type of generative model that utilizes an encoder-decoder
generative models, you can create new, synthetic data points that architecture. The encoder learns a lower-dimensional
can be added to an already existing dataset. This is typically used representation (latent space) of the input data and the decoder
in machine learning and deep learning applications to enhance rebuilds the input data from the latent space.
model performance, achieved by increasing both the size and the
VAEs force a probabilistic structure on the latent space that lets
diversity of the training data.
them create new data points by sampling from learned distribution.
Data augmentation can help to overcome challenges of imbalance These models are useful for data augmentation tasks with input
or limited datasets. By creating new data points similar to the data that has a complex structure, like text or images.
original data, data scientists can make sure that models are
stronger and better at generalizing unseen data. Generative Adversarial Networks (GANs)

Generative AI models like Variational Autoencoders (VAEs) and Consisting of two neural networks, a discriminator and a generator,
Generative Adversarial Networks (GANs) are promising for the that are simultaneously trained. The generator creates synthetic
generation of high-quality synthetic data. They learn the underlying data points and the discriminator assesses the quality of the
distribution of input data and are able to create new samples that created data by comparing it to the original data.
very closely resemble original data points Both the generator and the discriminator compete against each
other, with the generator attempting to create realistic data points
to deceive the discriminator. The discriminator tries to accurately
tell apart real and generated data, and as the training progresses,
the generator gets better at producing high-quality synthetic data.

39
Generative AI 2023 Report

There are several applications of generative AI data

augmentation models:
4. Time series analysis

1. Medical imaging Generating synthetic time series data by modeling underlying

patterns and creating new sequences with similar characteristics,
The generation of synthetic medical imaging like MRI scans or which can help enhance the performance of anomaly detection,
X-rays helps to increase the size of training datasets and enhance time series forecasting, and classification models.
diagnostic model performance.

5. Autonomous systems
2. Natural language processing (NLP)
Creating synthetic sensor data for autonomous vehicles and drones
Creating new text samples by changing existing sentences, allows the safe and extensive training of artificial intelligence
like replacing words with synonyms, adding noise, or changing systems without including real-world risks.
word order. This can help enhance the performance of machine
translation models, text classification, and sentiment analysis.
6. Robotics

3. Computer vision Generating both synthetic objects and scenes lets robots be trained
for tasks like navigation and manipulation in virtual environments
The enhancement of image datasets by creating new images with before they’re deployed into the real world.
different transformations, like translations, rotations, and scaling.
Can help to enhance the performance of object detection, image
classification, and segmentation models.

40
Generative AI 2023 Report

How do generative AI data augmentation models work?

Augmented data derives from original data with minor changes and synthetic data is artificially generated without using the
original dataset. The latter often uses GANs and deep neural networks (DNNs) in order to generate synthetic data.

There are a few data augmentation techniques:

Text data augmentation Audio data augmentation Image data augmentation

Sentence or word shuffling. Noise injection. Color space transformations.

Change the position of a sentence or Add random or Gaussian noise to audio Change the RGB color channels,
word randomly. datasets to enhance model performance. brightness, and contrast randomly.

Word replacement. Shifting. Image mixing.

You can replace words with synonyms. Shift the audio left or right with Blend and mix multiple images.
random seconds.

Syntax-tree manipulation. Changing speed. Geometric transformations.

Paraphrase the sentence by using the Stretches the times series by a fixed rate. Crop, zoom, flip, rotate, and stretch
same word. images randomly; however, be careful
when applying various transformations
on the same images, as it can reduce the
model’s performance.

Random word insertion. Changing pitch. Random erasing.

Add words at random. Change the audio pitch randomly. Remove part of the original image.

Random word deletion. Kernel filters.

Remove words at random. Change the blurring or sharpness of the
image randomly.

41
Generative AI 2023 Report

Visual/video applications
Generative AI is becoming increasingly important for video 3. Personalized content
applications due to its ability to produce, modify, and analyze video
content in ways that were previously impractical or impossible. Generative AI can change videos to fit individual preferences or
requirements. For example, a scene could be adjusted to show a
With the growing use of generative AI for video applications,
viewer’s name on a signboard, or a product that the viewer had
however, some ethical concerns arise. Deep Fakes, for example,
previously expressed interest in.
have been used in malicious ways, and there’s a growing need for
tools to detect and counteract them.
4. Virtual reality and gaming
Authenticity verification, informed consent for using someone’s
likeness, and potential impacts on jobs in the video production Generative AI can be used to generate realistic, interactive
industry are just some of the challenges that still need to be environments or characters. This offers the potential for more
navigated. dynamic and responsive worlds in games or virtual reality
experiences.
There are several applications of generative AI video models:

5. Training
1. Content creation

Due to its ability to create diverse and realistic scenarios,

Generative models can be used to create original video content,
generative AI is great for training purposes. It can generate various
such as animations, visual effects, or entire scenes. This is especially
road scenarios for driver training or medical scenarios for training
important for filmmakers or advertisers on a tight budget who
healthcare professionals.
might not be able to afford extensive CGI or live-action shoots.

6. Data augmentation
2. Video enhancement

For video-based machine learning projects, sometimes there

Generative models can upscale low-resolution videos to higher
isn’t enough data. Generative models can create additional video
resolutions, fill in missing frames to smooth out videos, or restore
data that’s similar but not identical to the existing dataset, which
old or damaged video footage.
enhances the robustness of the trained models.

42
Generative AI 2023 Report

7. Video compression

Generative models can help in executing more efficient video

compression techniques by learning to reproduce high-quality
videos from compressed representations.

8. Interactive content

Generative models can be used in interactive video installations or

experiences, where the video content responds to user inputs in real
time.

9. Marketing and advertising

Companies can use generative AI to create personalized video

ads for viewers or to quickly generate multiple versions of a video
advertisement for A/B testing.

10. Video synthesis from other inputs

Generative AI can produce video clips from textual descriptions

or other types of inputs, allowing for new ways of storytelling or
visualization techniques.

43
Generative AI 2023 Report

How do generative AI video models work?

Generative video models are computer programs that create new • Conditional generative models. These models create videos
videos based on existing ones. They learn from video collections based on specific given attributes or data. Factors like
and generate new videos that look both unique and realistic. computational needs, intricacy, and project-specific demands
need to be taken into account when selecting.
With practical applications in virtual reality, film, and video game
development, generative video models can be used for content
Training process for the video generation model
creation, video synthesis, and special effects generation.
The structure and hyperparameters for the selected generative
Creating a generative video model involves:
model are outlined. The curated video data teaches the model,
Preparing video data aiming to create both believable and varied video sequences.
The model’s efficacy needs to be checked consistently using the
The first step includes gathering a varied set of videos reflecting the
validation dataset.
kind of output to produce. Streamlining and refining this collection
by discarding any unrelated or subpar content guarantees both
Refining the output
quality and relevancy. The data must then be organized into
separate sets for training and validating the model’s performance. If needed, the generated sequences need to be adjusted to
uplift their clarity and continuity. Employ various enhancement
Choosing the right generative model techniques, such as diminishing noise, stabilizing the video, or
adjusting colors.
Picking an appropriate architecture for generating videos is vital.
Potential choices include Variational Autoencoders (VAEs) or
Assessment and optimization of the model
Generative Adversarial Networks (GANs). The options are:
The produced videos need to be examined by using multiple
• Variational Autoencoders (VAEs). These models acquire a latent
criteria, like their visual appeal, authenticity, and variety. Opinions
understanding of videos and then craft new sequences by
from specialized users or experts can be helpful in gauging the
pulling samples from this acquired latent domain.
utility and efficiency of the video-generating model.
• Generative Adversarial Networks (GANs). These models consist
of a generator and discriminator that work in tandem to Putting the model to use
produce lifelike videos. If everything is working as it should, the model can be launched to
• Recurrent Neural Networks (RNNs). Models adept at recognizing produce new video sequences. The video generation model can be
time-based patterns in videos, producing sequences grounded utilized in diverse areas, including video creation, special cinematic
in these identified patterns. effects, or immersive experiences in virtual reality.

44
Section 4

Main generative AI tool of

choice, benefits, and challenges

45
Generative AI 2023 Report

Main generative AI tool of choice, benefits, and challenges What is the main benefit of your number one choice?

We asked respondents to tell us what the main benefit was of their

Bard 5.1% number one choice of generative AI tool. Perhaps a bit surprisingly,
cost savings was the least mentioned benefit (1.2%), with efficiency
being the number one benefit (26.7%).
ChatGPT 78.3%

Speed 15.5%
Midjourney 3.2%

Accuracy 10.3%
GitHub Copilot 2.3%

Quality 14.4%

Other 11.1%

Efficiency 26.7%

Cost savings 1.2%

Better decision-making 8.2%

Increased creativity 15.5%

Other 8.2%

Let’s have a look at what the benefits can entail:

46
Generative AI 2023 Report

Speed Quality
The ability to generate data, designs, or content at rapid speeds is “Quality over quantity” gets a new life with generative AI. While
essential in today’s world. Generative AI stands out as a tool that the technology can produce large volumes of output, it doesn’t
can meet this demand, acting as an engine that can produce necessarily compromise the quality of the results. As AI models
results far quicker than conventional methods. evolve and learn, they aim to generate outputs of high quality.

As it can create large volumes of data in no time, like text, images, In addition to producing better-quality outputs, generative AI
and more, it can have an impact in a wide variety of fields. From can also learn and get feedback over time, meaning it can keep
writing articles to generating music and more, these tools can improving itself and offer even better results in the future.
speed up employee productivity.
Efficiency
They also help in designing and manufacturing numerous design
concepts or variants, which greatly reduces the time that’s With generative AI, processes are streamlined and optimized,
traditionally needed. ensuring that resources, both time and computational, are used to
their fullest potential. This means that maximum productivity with
Accuracy minimum wasted effort or expenses can be achieved.

Precision is essential in many sectors, from medical diagnoses to The technology can operate with minimal human oversight, which
financial forecasting. Generative AI is designed to identify intricate frees up resources. And it can be integrated into workflows, leading
patterns and produce results that match these patterns, minimizing to automated processes that make the overall systems more
the margin of error and enhancing the dependability of outputs. efficient.

Operating on algorithms and training data, generative AI is

Cost savings
consistent in its outputs without typical human errors. The tools can
also be trained to produce outputs that closely match real-world One of the main goals of any business is to reduce costs without
patterns by learning from extensive datasets. compromising on results. By automating and refining processes,
generative AI tools can be a big investment in financial stability,
resulting in substantial savings long term.

As generative AI can automate tasks, there’s a decreased need for

manual hours and higher cost savings. This automation also leads
to outputs of higher accuracy and quality, needing fewer revisions
and saving time and money.

47
Generative AI 2023 Report

Better decision-making We also took a closer look at the top generative AI tools and how
they were seen according to the top mentioned benefits:
Particularly when it comes to complex scenarios, it can be difficult
to make decisions. Generative AI can guide companies with their
decision-making processes, providing insights, simulations, and
Speed
forecasts that can open the way for more informed and strategic
Efficiency 20.6%
choices.
28.6%
Having data-driven insights helps to simulate multiple scenarios
or even create a varied output – this lets decision-makers have
a comprehensive view of every possibility. Predictive analysis can
foresee potential outcomes based on historical data to guide
decisions. Increased
creativity
Increased creativity 17.5%
Quality
Generative AI can act as a catalyst for increased creativity. By 33.3%
combining patterns in innovative ways, artificial intelligence tools
can open the door to groundbreaking ideas and creations. This
technology can also inspire creators to think outside the box or view
problems from different angles. Accuracy
50%
Other
Better
Respondents who chose “other” added a few different benefits:
decision-making
• Conversational interface that allows me to better describe the 16.7%
initial question and go deeper into any provided answers
• Sources
• Availability
• Easy to use Increased
creativity
• Responsible enterprises grade
17.5%
• Most popular
Increased
• Convenience creativity
66.7%

48
Generative AI 2023 Report

What is the main challenge of your number one choice?

We wanted to know what was the main challenge with respondents’ number one generative AI tool.

As you can see below, biases, errors, and limitations of generative AI were highlighted as the main concern when using the
technology (36.1%), followed by a limited information pool (19.5%) and generative AI data security concerns (16.5%) in second and
third place, respectively

Generative AI data security 16.5%

Generative AI vs. IP rights 5.2%

Biases, errors, and limitations of generative AI 36.1%

Dependence on the 3rd party platform 2.1%

Limited information pool 19.5%

Costs associated with premium versions 12.4%

Workforce morale 1%

Other 7.2%

49
Generative AI 2023 Report

Now, let’s take a closer look at each of these challenges:

Generative AI data security

Generative AI vs. IP rights
When generative models are trained on vast datasets,
they can unintentionally memorize personal or sensitive As generative AI creates artworks, written content, and
information. And, when generating content for other users, music pieces, the question arises: who owns the rights to this
there’s a possibility that models will leak this information. content? The AI developer, the user? This can be a big point of
contention when using the models.
Models can also create false data or manipulate existing
data, which presents challenges in guaranteeing data Similarly, generative AI can produce content that is very close
integrity – particularly in critical sectors like finance or to existing copyrighted work and lead to legal disputes, or
healthcare. And, generative AI models can be susceptible to even recreate products or content based on existing ones and
adversarial attacks, a slight deviation from the input can lead undermine the value of the original creations.
to models producing incorrect outputs.

Biases, errors, and limitations of generative AI Dependence on the 3rd party platform

Should the data used to train models have biases, the There can be an over-reliance on specific platforms,
created content will reflect these. This can be a big concern meaning that migrating to other platforms or solutions
when it comes to real-world applications in law enforcement, becomes difficult. Should these third-party platforms have
hiring processes, and healthcare. outages or discontinuations, users who depend on them can
face serious disruptions.
There are also ethical concerns when it comes to generative
AI, as it can create or amplify harmful stereotypes. The model Relying on these external platforms can result in databeing
outputs can be unpredictable at times, and this can lead to stored in locations that are subject to different law regulations
errors or even impropriate content generation. – or even to data being accessed by unwanted parties.

50
Generative AI 2023 Report

Costs associated with premium versions

Limited information pool
Premium versions of generative AI platforms can often have
Generative AI models are only as good as the data used to
additional or advanced features that make work simpler and
train them; if they’re trained on limited datasets, outputs can
more efficient. They can, however, be expensive, which places
lack diversity or be highly inaccurate. Additionally, artificial
them out of reach for small businesses or individuals.
intelligence models have a “knowledge cutoff”, meaning that
Alongside premium versions, there can also be hidden
they aren’t aware of new information or current events unless
costs associated with data storage, computation, or other
consistently updated.
infrastructure needs.

Other

Workforce morale Respondents also mentioned a few other challenges that

their first generative AI tool of choice has. A few of these are:
With the quick adoption of AI, there’s been a feat that it could
replace certain job roles, especially generative AI, leading to • Its speed
reduced job security or unemployment. Employees might also • Its availability
be wary of AI-generated content and doubt its reliability and • Lack of speech recognition
accuracy.
• How hard it is to find someone who knows how to use the
As this adoption grows, so does the need for upskilling. tools well enough to get the most out of them.
Employees could feel overwhelmed or even left behind if they • Knowing about what’s out there since there are so many
aren’t provided with adequate training. tools available.
• The cost of initial development is more than expected.

51
Section 5

How much has your generative

AI tool of choice impacted your
productivity and efficiency?

52
Generative AI 2023 Report

How much has your generative AI tool of choice impacted your productivity and efficiency?
Alongside knowing the benefits and drawbacks of generative AI tools, we also set out to find out how respondents feel about the
impact the technology has on their productivity and efficiency.

Positive impact Negative impact

0 3.1% 0 45.3%

10% 4.1% 10% 9.3%

20% 0% 20% 19.5%

30% 3.1% 30% 8.2%

40% 5.2% 40% 2.2%

50% 9.3% 50% 3.1%

60% 12.4% 60% 4.1%

70% 17.5% 70% 2.1%

80% 20.6% 80% 5.2%

90% 4.1% 90% 1%

100% 20.6% 100% 0%

Average total of positive productivity impact: 68.6%. Average total of negative productivity impact: 9.8%.

62.9% said that generative AI tools positively affect productivity and Only 8.2% said that generative AI tools negatively impact
efficiency at 70% or higher. productivity and efficiency at 70% or higher.

53
Section 6

Trust and quality

54
Generative AI 2023 Report

We wanted to find our how much respondents trust generative AI tools and how they rate them in terms of the quality
of content created.

How much do you trust the generative AI tools?

With the large majority of respondents using generative AI tools, it’s perhaps surprising to find that only 38.5% trust it
at 7 or above. It’s interesting to reflect that low trust hasn’t impeded the use of the technology.

0 1 2 3 4 5 6 7 8 9 10
3.1% 4.2% 4.2% 7.3% 8.3% 22.9% 11.5% 13.5% 13.5% 6.3% 5.2%

Average total of how much respondents trust generative AI tools: 53.1%.

55
Generative AI 2023 Report

How would you rate the quality of the generated content?

On average, the majority of respondents rated their generative AI number one choice at a four out of five (48.5%).

The average rating of the quality of all tools sits at a somewhat comfortable 3.9 stars out of 5 – perhaps higher than expected
when we analyze the little trust that respondents seem to have in the tools.

Total average rating: 3.9

We selected the top three rated generative AI tools mentioned, ChatGPT, Google Bard, and Midjourney, and evaluated how
respondents rate the quality of produced content.

ChatGPT

Total average rating: 3.8

Bard

Total average rating: 4

Midjourney

Total average rating: 4

56
Section 7

Non-use of generative AI tools

57
Generative AI 2023 Report

11.8% of respondents said they don’t use generative AI tools, so we wanted to know what was behind this choice.

Why have you chosen not to use generative AI tools?

The majority of respondents, 46.2%, identified lack of knowledge as the number one reason for not using generative AI
tools. This points to a need for companies to train employees in these evolving technologies.

Maybe unsurprisingly, the second most chosen reason for not using the tools was data privacy concerns with 30.8% of
votes, followed by concerns over IP with 15.4%.

Other
7.7%

Lack of knowledge
46.2%

Data privacy
30.8%

Concerns over IP
15.4%

58
Generative AI 2023 Report

Let’s have a closer look at the reasons mentioned above as to why respondents don’t use generative AI tools:

Lack of knowledge
Concerns over IP
Highlighted as the number one reason for the lack of
generative AI adoption, a lack of knowledge about generative Copyright laws are complicated and the legal status of
AI tools can be a big blocker for companies. generative AI could take quite a while to settle. Generative
models like ChatGPT and DALL-E make new patterns from
Generative AI can be complex, often involving intricate
millions of examples in training sets, meaning that their results
algorithms and deep learning models. Without proper
are derived from various existing sources.
understanding, it can be difficult for companies to correctly
integrate, use, and derive value from these models. When these tools generate content, the data’s source could be
unknown; companies in finance, for example, could be relying
Additionally, employees also need to understand what
on another company’s intellectual property. This could lead to
generative AI can and can’t do, as otherwise companies might
reputational and financial risks.
have misaligned expectations. The tools’ capabilities can be
underestimated, leading companies to miss out on potential There need to be legal precedents offering clarity around both
applications, for example, or even disappointment if the IP and copyright challenges when it comes to generative AI.
technology doesn’t match expectations.

Data privacy

Large language models (LLMs) are trained on vast datasets that can include personally identifiable information. This data can be
obtained with a simple text prompt, and unlike traditional search engines, it can be considerably more difficult for individuals to locate
and request the removal of this information.

So, companies building or upgrading LLMs need to make sure that personally identifiable information isn’t embedded in the language
models. They also need to make sure that this information is easy to remove from the models, according to privacy laws.

59
Generative AI 2023 Report

Would you consider using generative AI tools The majority responded that their companies view generative AI
in the future? in a neutral way, quickly followed by a positive opinion. Maybe
somewhat surprisingly, no one said their company was against the
100% of respondents said they’d consider using generative AI tools
technology.
in the future, and perhaps unsurprisingly, the vast majority (76.9%)
said they’d consider using ChatGPT.
Do you trust generative AI tools?
We wanted to know how much those who don’t use the technology
Yes
trust it. In a perhaps surprising result, not everyone who said they’d
100%
use the technology in the future also said they trusted it.

No
27.2%

Yes
72.8%

The general consensus about the use of generative AI

We wanted to know what the opinion was in the respondents’
companies about generative AI.

For - 46.2%

Against - 0%

Neutral - 53.8%

60
Section 8

Generative AI future insight

61
Generative AI 2023 Report

Paul Claxton
AI Venture Capitalist, Serial Entrepreneur

How do you think generative AI technology will evolve in the next 5 years?

As businesses seek to gain consumer insights, generative AI As the dependency on AI continues to rise and humans become
has emerged as a critical tool. In the next five years, I envision increasingly reliant on generative AI, I believe that without the
generative AI becoming the primary interface of internet search development of complementary technologies, we may be headed
engines, allowing for a more natural and intuitive user experience. toward a type of stagflation. This would result in rising costs, such
as labor, but without corresponding opportunities to increase
Today, generative AI functions reactively, relying on inputs from
prices. This would lead to displacement, inflation, and ultimately
humans to generate text, images, music, video, and code. However,
slow economic growth, as a result of high demand and low supply
as usage increases, I believe it will evolve into a more proactive
of quality AI systems, expertise, and real-world solutions that are
system. This growth will result in what I call “Administrative AI,”
applicable and economically impactful.
where the AI inputs to humans, and humans respond and answer
back to the AI system. If we neglect the migration to multimodal large language models
from large language models and the implementation of essential
Although generative AI is data-centric, its progress is limited.
areas integral to AI, such as cybersecurity, computer vision/Lidar,
Advancements in multimodal large language models and other
edge computing, and quantum computing, the progression of
critical areas like computer vision/Lidar technologies, edge
generative AI could slow down.
computing, quantum computing, and cybersecurity are necessary
to improve generative AI performance. To tap into the full potential of the technology and maximize
its benefits, we must ensure that the component technologies
In the next five years, I foresee further development in these
necessary to drive it forward are in place. While generative AI
areas that will integrate generative AI into our daily lives. For
currently relies on human input for creativity, significant progress in
example, instead of using generative AI just on our laptops and
component technologies could unleash groundbreaking use cases
phones, the technology could be built into refrigerators, allowing
for the future during the next 5 years.
them to visually scan the contents, learn buying habits, and alert
homeowners when items are running low.

62
Generative AI 2023 Report

What do you think generative AI will impact the most?

Generative AI has the potential to impact various industries but AI is being built to be objective for a subjective population, and
to determine where it will have the most impact, we should look it can help humans avoid bias and misinformed decisions. The
beyond just tangible industries and focus on the intangible global objective of AI is to create a healthy society by obtaining accurate
industry of social determinants of healthcare. guidance and perspective so that it can help humans navigate and
deal with threats to society and humanity, such as COVID, or even
While social determinants of healthcare aren’t a traditional
with ourselves.
industry, it is an area where AI can make a significant impact. The
sociological factors that have long affected every industry will either Therefore, when we talk about AI and its impact on industries,
be accentuated or deteriorated by AI. we must consider where it will have the greatest impact to help
humans evolve and become better in every industry and every
Generative AI is likely to have the greatest impact in opaque
way. For any and all impactful advancements, everything comes
and abstract areas, such as healthcare. Specifically, it can help
down to ethics, when we are thinking about the social determinants
with culturally, socially, and medically significant issues, such as
of healthcare, ethics will continue to play a greater part of the AI
COVID-19, which impacts healthcare not only in the medical sense
revolution in considering what is real and what is not, and the power
but also in the existential sense.
we derive from AI as well as bestow upon it.
Humans are now reliant on AI, and generative AI has the potential to
***These are only my opinions based on my experience as a Marine,
impact human sociology for every industry, which is a dimension of
Venture Capitalist, Serial Entrepreneur, Consumer of AI, and Global
health and wellness. Sociology is more opaque than even globally
Citizen.***
borne plagues, and therefore, generative AI has the potential to
make the greatest long-term impact involving social determinants
of health. It can help with bias reduction and provide increased
understanding for every industry.

63
Generative AI 2023 Report

Srikanth Mahankali, MD
Founder & CEO (Shree Advisory & Consulting, LLC)

How do you think generative AI technology will evolve in the next 5 years?

I see it having wide application in everyday life going forward. From just being a source of information all the way to even being a
“therapist”. There are of course issues with accuracy, IP, and regulation that will need to be overcome before it can be labeled trustworthy
and dependable.

Affixing liability based on actions taken from feedback by generative AI technology is another area that requires significant attention.
Overall… the promise and potential are there and how it will evolve over the next five years is entirely dependent on how it is used!

What do you think generative AI will impact the most?

I can see it having a significant impact in healthcare as a source of information, preparing patients before a physician visit, and the like. I
have also become aware that physicians (particularly the technology-savvy ones) start to use it regularly in their daily work as a source of
quick reference, write letters to patients, prepare discharge summaries, and the like.

Another domain that I think generative AI will potentially help immensely in is mental health as a source of conversation and empathy to
combat loneliness, isolation, and lack of caregiver support. Outside of healthcare - I can see it bringing great value to the hospitality and
financial industries.

64
Generative AI 2023 Report

Raghavendra Rao Gudipudi Vishnuprasanth Vijaya Kumar

AmberFlux EdgeAI Cigna-Evernorth

How do you think generative AI technology will How do you think generative AI technology will
evolve in the next 5 years? evolve in the next 5 years?

Most enterprise applications will have some component of Human life will speed up, behaviors will change and industries will
generative AI. Generative AI will become more specific and be transformed and that’s what can be predicted with certainty.
generative AI tools for specific business functions will proliferate.

What do you think generative AI will impact What do you think generative AI will impact
the most? the most?

It will revolutionize healthcare as both Hospitals and Patients have Generative AI will have a significant impact across all industry
better tools for communication. For example, doctors will have a sectors. Banking, high tech, and life sciences are among the
chance to get immediate diagnostic results using AI tools, and this industries that could see the biggest impact as a percentage of
will enable a reduced number of visits or amount of time spent their revenues from generative AI.
by the patient. For a patient instead of calling the doctor, he can
converse with generative AI tools to get all his questions answered.

65
Generative AI 2023 Report

Sahiz Kaur
White Cup

How do you think generative AI technology will

evolve in the next 5 years?

We will have solutions with specific knowledge bases. These will act
as the onboarding tools to get new people up to speed.

What do you think generative AI will impact

the most?

I don’t know how, but I believe there have to be consequences of us

outsourcing our research and writing work to generative AI. I don’t
believe it will show up soon enough for our generation, however, it
will be interesting to see how the next generations respond to it and
whether or not it has a significant impact on brain growth. I believe
that will inevitably lead to changes in the healthcare industry
because we will have a new problem to solve and train our medical
staff to treat.

66
Generative AI 2023 Report

Conclusion
TL;DR
No time to read the full report just yet?

No worries, we’ve got you covered with the highlights!

• Companies on opposite ends of the revenue spectrum (under $50 million and over
$500 million) both use generative AI due to its efficiency and both have the same
concerns (biases, errors, and limitations of generative AI).

• They both also use this technology mostly to speed up work.

• ChatGPT is the most used tool.

• Efficiency is the number one advantage of using generative AI and biases, errors, and
limitations of generative AI are the biggest concerns.

• Of the respondents who don’t use generative AI tools, all said they’d consider using
them in the future; however, a small percentage claimed to not trust them.

• On average, only 53.1% of respondents who use generative AI trust the technology.

• A fifth of the respondents say generative AI doubled their productivity.

67
Generative AI 2023 Report

Contributors
We would like to say a special thank you to all report contributors for their insights and
expertise. This report wouldn’t have been possible without them.

Marisa Garanhel
Copywriter
Marisa is our Copywriter here at AIAI who loves crafting content to keep readers
informed, entertained, and enthralled. She’s always open to feedback and
would be thrilled to hear from you!

Jordan Dunne
Co-Founder, AIAI
Jordan, Co-Founder AIAI, is a firm believer in a data-driven future led by AI & is
passionate about connecting AI experts and machine scientists from around
the world.

Lucy Sproule
Senior Graphic Designer
Meet Lucy, our talented Graphic Designer who specializes in fulfilling all
our design needs. She's in charge of creating the layout and incorporating
visual elements in this report and is always eager to hear your thoughts
on her designs!

68
Generative AI 2023 Report

Contributors

Paul Claxton,
AI Venture Capitalist, Serial Entrepreneur

Srikanth Mahankali,
MD, Founder & CEO (Shree Advisory & Consulting, LLC)

Raghavendra Rao Gudipudi,

AmberFlux EdgeAI

Vishnuprasanth Vijaya Kumar,

Cigna-Evernorth

Sahiz Kaur,
White Cup

69