Artificial Intelligence

Index Report 2023

Artificial Intelligence
Index Report 2023

CHAPTER 4:
The Economy

Chapter 4 Preview 1
 Artificial Intelligence
Index Report 2023

CHAPTER 4 PREVIEW:

The Economy
Overview 3 Narrative Highlight: The Effects of
Chapter Highlights 4 GitHub’s Copilot on Developer
Productivity and Happiness 41

4.1 Jobs 6 Industry Motivation 43

AI Labor Demand 6 Perceived Importance of AI 43

Global AI Labor Demand 6 AI Investments and Investment

Outcomes 44
U.S. AI Labor Demand by Skill Cluster
and Specialized Skill 7 Challenges in Starting and Scaling
AI Projects 46
U.S. AI Labor Demand by Sector 9
Earnings Calls 48
U.S. AI Labor Demand by State 10
Aggregate Trends 48
AI Hiring 13
Specific Themes 49
AI Skill Penetration 15
Narrative Highlight: What Are Business
Global Comparison: Aggregate 15
Leaders Actually Saying About AI? 50
Global Comparison: By Gender 16
Sentiment Analysis 52

4.2 Investment 17
4.4 Robot Installations 53
Corporate Investment 17
Aggregate Trends 53
Startup Activity 20
Industrial Robots:
Global Trend 20 Traditional Vs. Collaborative Robots 55
Regional Comparison by Funding Amount 22 By Geographic Area 56
Regional Comparison by Narrative Highlight: Country-Level
Newly Funded AI Companies 26 Data on Service Robotics 60
Focus Area Analysis 28 Sectors and Application Types 63
China Vs. United States 65
4.3 Corporate Activity 31
Industry Adoption 31 Appendix 67

Adoption of AI Capabilities 31
Consideration and Mitigation of Risks ACCESS THE PUBLIC DATA
From Adopting AI 39

Chapter 4 Preview 2
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023

Overview
Increases in the technical capabilities of AI systems have led to greater rates of AI
deployment in businesses, governments, and other organizations. The heightening
integration of AI and the economy comes with both excitement and concern. Will
AI increase productivity or be a dud? Will it boost wages or lead to the widespread
replacement of workers? To what degree are businesses embracing new AI
technologies and willing to hire AI-skilled workers? How has investment in AI
changed over time, and what particular industries, regions, and fields of AI have
attracted the greatest amount of investor interest?

This chapter examines AI-related economic trends by using data from Lightcast,
LinkedIn, McKinsey, Deloitte, and NetBase Quid, as well as the International
Federation of Robotics (IFR). This chapter begins by looking at data on AI-related
occupations and then moves on to analyses of AI investment, corporate adoption of
AI, and robot installations.

Chapter 4 Preview 3
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023

Chapter Highlights
The demand for AI-related For the first time in the last
professional skills is increasing decade, year-over-year private
across virtually every investment in AI decreased.
American industrial sector. Global AI private investment was $91.9 billion
Across every sector in the United States for in 2022, which represented a 26.7% decrease
which there is data (with the exception of since 2021. The total number of AI-related
agriculture, forestry, fishing, and hunting), the funding events as well as the number of newly
number of AI-related job postings has increased funded AI companies likewise decreased.
on average from 1.7% in 2021 to 1.9% in 2022. Still, during the last decade as a whole, AI
Employers in the United States are increasingly investment has significantly increased. In 2022
looking for workers with AI-related skills. the amount of private investment in AI was 18
times greater than it was in 2013.

Once again, the United States leads in investment in AI.

The U.S. led the world in terms of total amount of AI private investment. In 2022, the $47.4 billion
invested in the U.S. was roughly 3.5 times the amount invested in the next highest country, China
($13.4 billion). The U.S. also continues to lead in terms of total number of newly funded AI companies,
seeing 1.9 times more than the European Union and the United Kingdom combined, and 3.4 times
more than China.

In 2022, the AI focus area with the most investment was medical
and healthcare ($6.1 billion); followed by data management,
processing, and cloud ($5.9 billion); and Fintech ($5.5 billion).
However, mirroring the broader trend in AI private investment, most AI focus areas saw less investment
in 2022 than in 2021. In the last year, the three largest AI private investment events were: (1) a $2.5 billion
funding event for GAC Aion New Energy Automobile, a Chinese manufacturer of electric vehicles; (2) a
$1.5 billion Series E funding round for Anduril Industries, a U.S. defense products company that builds
technology for military agencies and border surveillance; and (3) a $1.2 billion investment in Celonis, a
business-data consulting company based in Germany.

Chapter 4 Preview 4
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023

Chapter Highlights (cont’d)

While the proportion of AI is being deployed
companies adopting AI has by businesses in
plateaued, the companies multifaceted ways.
The AI capabilities most likely to have been
that have adopted AI
embedded in businesses include robotic
continue to pull ahead. process automation (39%), computer vision
The proportion of companies adopting AI
(34%), NL text understanding (33%), and virtual
in 2022 has more than doubled since 2017,
agents (33%). Moreover, the most commonly
though it has plateaued in recent years
adopted AI use case in 2022 was service
between 50% and 60%, according to the
operations optimization (24%), followed by
results of McKinsey’s annual research
the creation of new AI-based products (20%),
survey. Organizations that have adopted AI
customer segmentation (19%), customer
report realizing meaningful cost decreases
service analytics (19%), and new AI-based
and revenue increases.
enhancement of products (19%).

AI tools like Copilot are China dominates industrial

tangibly helping workers. robot installations.
Results of a GitHub survey on the use of In 2013, China overtook Japan as the nation
Copilot, a text-to-code AI system, find installing the most industrial robots. Since
that 88% of surveyed respondents feel then, the gap between the total number of
more productive when using the system, industrial robots installed by China and the
74% feel they are able to focus on more next-nearest nation has widened. In 2021,
satisfying work, and 88% feel they are able China installed more industrial robots than
to complete tasks more quickly. the rest of the world combined.

Chapter 4 Preview 5
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.1 Jobs

4.1 Jobs
AI Labor Demand Global AI Labor Demand
Figure 4.1.1 highlights the percentage of all job
This section reports demand for AI-related skills
postings that require some kind of AI skill. In 2022,
in labor markets. The data comes from Lightcast,
the top three countries according to this metric were
which mined millions of job postings collected from
the United States (2.1%), Canada (1.5%), and Spain
over 51,000 websites since 2010 and flagged listings
(1.3%). For every country included in the sample, the
calling for AI skills.
number of AI-related job postings was higher in 2022
than in 2014.1

AI Job Postings (% of All Job Postings) by Geographic Area, 2014–22

Source: Lightcast, 2022 | Chart: 2023 AI Index Report

2.05%, United States

2.00%
AI Job Postings (% of All Job Postings)

1.45%, Canada
1.50% 1.33%, Spain
1.23%, Australia
1.20%, Sweden
1.16%, Switzerland
1.14%, United Kingdom
1.01%, Netherlands
0.98%, Germany
1.00% 0.89%, Austria
0.86%, Belgium
0.84%, France
0.72%, Italy

0.50%
0.45%, New Zealand

0.00%
2014 2015 2016 2017 2018 2019 2020 2021 2022

Figure 4.1.1

1 In 2022, Lightcast slightly changed their methodology for determining AI-related job postings from that which was used in previous versions of the AI Index Report. As such, some of the
numbers in this chart do not completely align with those featured in last year’s report.

Chapter 4 Preview 6
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.1 Jobs

U.S. AI Labor Demand by Skill Cluster and Specialized Skill

Figure 4.1.2 showcases the most in-demand AI skill clusters in the U.S. labor market since 2010. The most
in-demand skill cluster was machine learning (1.0%), followed by artificial intelligence (0.6%) and natural
language processing (0.2%). Every listed AI skill cluster is now more in demand than it was 10 years ago.

AI Job Postings (% of All Job Postings) in the United States by Skill Cluster, 2010–22
Source: Lightcast, 2022 | Chart: 2023 AI Index Report

1.03%, Machine Learning

1.00%
AI Job Postings (% of All Job Postings)

0.80%

0.60% 0.61%, Arti cial Intelligence

0.40%

0.20%, Natural Language Processing

0.20% 0.16%, Neural Networks
0.15%, Autonomous Driving
0.13%, Visual Image Recognition

0.06%, Robotics
0.00%

2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022

Figure 4.1.2

Chapter 4 Preview 7
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.1 Jobs

Figures 4.1.3 and 4.1.4 showcase the top ten specialized skills that were demanded in AI job postings in 2022 compared
to 2010–20122. On an absolute level, virtually every specialized skill is more in demand now than a decade ago. The
growth in demand for Python is particularly notable, evidence of its growing popularity as an AI coding language.

Top Ten Specialized Skills in 2022 AI Job Postings in the United States, 2010–12 Vs. 2022
Source: Lightcast, 2022 | Chart: 2023 AI Index Report
296,662
Python (Programming Language)
12,884
260,333
Computer Science
48,001
185,807
SQL (Programming Language)
22,037
159,801
Data Analysis
16,571
157,855
Data Science
1,227
155,615
Amazon Web Services
962
152,956
Agile Methodology
7,549
138,791
Automation
13,207
133,856
Java (Programming Language)
26,557 2022
133,286 2010–12
Software Engineering
22,384
0 50,000 100,000 150,000 200,000 250,000 300,000
Number of AI Job Postings Figure 4.1.3

Top Ten Specialized Skills in 2022 AI Job Postings in the United States by Skill Share, 2010–12 Vs. 2022
Source: Lightcast, 2022 | Chart: 2023 AI Index Report
37.13% (+592%)
Python (Programming Language)
5.36%
32.58% (+63%)
Computer Science
19.98%
23.25% (+153%)
SQL (Programming Language)
9.17%
20.00% (+190%)
Data Analysis
6.90%
19.75% (+3,767%)
Data Science
0.51%
19.47% (+4,763%)
Amazon Web Services
0.40%
19.14% (+509%)
Agile Methodology
3.14%
17.37% (+216%)
Automation
5.50%
16.75% (+52%)
Java (Programming Language)
11.06% 2022
16.68% (+79%) 2010–12
Software Engineering
9.32%
0% 5% 10% 15% 20% 25% 30% 35% 40%
Skill Share in AI Job Postings (%)
Figure 4.1.4
2 The point of comparison of 2010–2012 was selected because some data at the jobs/skills level is quite sparse in earlier years. Lightcast therefore used the
whole set of years 2010–2012 to get a larger sample size for a benchmark from 10 years ago to compare.

Chapter 4 Preview 8
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.1 Jobs

U.S. AI Labor Demand by Sector fishing, and hunting), the number of AI job postings
Figure 4.1.5 shows the percentage of U.S. job was notably higher in 2022 than in 2021, with the top
postings that required AI skills by industry sector three sectors being information (5.3%); professional,
from 2021 to 2022. Across virtually every included scientific, and technical services (4.1%); and finance
sector (with the exception of agriculture, forestry, and insurance (3.3%).

AI Job Postings (% of All Job Postings) in the United States by Sector, 2021 Vs. 2022
Source: Lightcast, 2022 | Chart: 2023 AI Index Report
5.30%
Information 4.85%
4.07%
Professional, Scienti c, and Technical Services 3.86%
3.33%
Finance and Insurance 2.94%
3.26%
Manufacturing 2.86%
1.64%
Agriculture, Forestry, Fishing, and Hunting 1.66%
1.53%
Educational Services 1.41%
1.37%
Management of Companies and Enterprises 1.08%
1.32%
Public Administration 0.98%
1.28%
Retail Trade 0.82%
1.27%
Utilities 1.10%
1.19%
Mining, Quarrying, and Oil and Gas Extraction 1.00%
0.98%
Wholesale Trade 0.82%
0.89%
Real Estate and Rental and Leasing 0.65%
2022
0.67%
Transportation and Warehousing 0.59% 2021
0.58%
Waste Management and Administrative Support Services 0.56%

0% 1% 2% 3% 4% 5%
AI Job Postings (% of All Job Postings)
Figure 4.1.5

Chapter 4 Preview 9
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.1 Jobs

U.S. AI Labor Demand by State Number of AI Job Postings in the United States by State, 2022
Source: Lightcast, 2022 | Chart: 2023 AI Index Report
Figure 4.1.6 highlights the number
of AI job postings in the United AK ME
970 2,227
States by state. The top three
VT NH MA
states in terms of postings were 1,571 2,719 34,603
California (142,154), followed by WA MT ND SD MN WI MI NY CT RI
31,284 833 1,227 2,195 11,808 8,879 25,366 43,899 8,960 2,965
Texas (66,624) and New York
(43,899). OR ID WY NE IA IL IN OH PA NJ
10,811 6,109 769 4,032 5,670 31,569 9,247 19,208 20,397 23,447

CA NV UT CO KS MO KY WV DC MD DE
142,154 6,813 6,885 20,421 7,683 10,990 4,536 887 9,606 16,769 3,503

AZ NM OK AR TN VA NC
19,514 3,357 5,719 7,247 11,173 34,221 23,854

TX LA MS AL GA SC
66,624 4,806 2,548 7,866 26,620 4,928

HI FL
2,550 33,585

Figure 4.1.6

Figure 4.1.7 demonstrates what Percentage of U.S. States Job Postings in AI, 2022
Source: Lightcast, 2022 | Chart: 2023 AI Index Report
percentage of a state’s total job
postings were AI-related. The top AK ME
0.88% 1.64%
states according to this metric
VT NH MA
were the District of Columbia 1.34% 1.20% 2.26%
(3.0%), followed by Delaware WA MT ND SD MN WI MI NY CT RI
2.48% 0.72% 1.04% 1.83% 1.22% 0.90% 1.77% 2.07% 1.66% 1.84%
(2.7%), Washington (2.5%), and
Virginia (2.4%). OR ID WY NE IA IL IN OH PA NJ
1.43% 1.89% 1.18% 1.14% 1.14% 1.63% 0.88% 1.07% 1.30% 2.04%
CA NV UT CO KS MO KY WV DC MD DE
2.21% 1.23% 1.54% 1.46% 1.43% 1.15% 0.85% 0.99% 2.95% 1.96% 2.66%
AZ NM OK AR TN VA NC
1.40% 1.36% 1.07% 2.03% 1.11% 2.42% 1.44%
TX LA MS AL GA SC
1.52% 0.87% 1.15% 1.31% 1.64% 0.87%
HI FL
1.46% 1%

Figure 4.1.7

Chapter 4 Preview 10
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.1 Jobs

Which states had the greatest Percentage of United States AI Job Postings by State, 2022
Source: Lightcast, 2022 | Chart: 2023 AI Index Report
share of AI job postings as a
share of all AI job postings in AK ME
0.12% 0.28%
the U.S. in 2022? California was
VT NH MA
first: Last year 17.9% of all AI job 0.20% 0.34% 4.35%
postings in the United States WA MT ND SD MN WI MI NY CT RI
3.93% 0.10% 0.15% 0.28% 1.48% 1.12% 3.19% 5.52% 1.13% 0.37%
were for jobs based in California,
followed by Texas (8.4%) and OR ID WY NE IA IL IN OH PA NJ
1.36% 0.77% 0.10% 0.51% 0.71% 3.97% 1.16% 2.41% 2.56% 2.95%
New York (5.5%) (Figure 4.1.8). CA NV UT CO KS MO KY WV DC MD DE
17.87% 0.86% 0.87% 2.57% 0.97% 1.38% 0.57% 0.11% 1.21% 2.11% 0.44%
AZ NM OK AR TN VA NC
2.45% 0.42% 0.72% 0.91% 1.40% 4.30% 3%
TX LA MS AL GA SC
8.37% 0.60% 0.32% 0.99% 3.35% 0.62%
HI FL
0.32% 4.22%
Figure 4.1.8

Figure 4.1.9 highlights the trends over time in AI job postings for four select states that annually report a high
number of AI-related jobs: Washington, California, New York, and Texas. For all four, there was a significant
increase in the number of total AI-related job postings from 2021 to 2022, suggesting that across these states,
employers are increasingly looking for AI-related workers.

Percentage of U.S. States’ Job Postings in AI by Select U.S. State, 2010–22

Source: Lightcast, 2022 | Chart: 2023 AI Index Report

2.50% 2.48%, Washington

2.21%, California
Percentage of U.S. States’ Job Postings in AI

2.07%, New York

2.00%

1.50% 1.52%, Texas

1.00%

0.50%

0.00%
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Figure 4.1.9

Chapter 4 Preview 11
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.1 Jobs

Figure 4.1.10 highlights the degree to which AI-related job postings have been subdivided among the top
four states over time. California’s share of all AI job postings has decreased steadily since 2019 while Texas’
has marginally increased. The fact that California no longer commands one-quarter of all AI-related jobs
suggests that AI jobs are becoming more equally distributed among U.S. states.

Percentage of United States AI Job Postings by Select U.S. State, 2010–22

Source: Lightcast, 2022 | Chart: 2023 AI Index Report

25%
Percentage of United States AI Job Postings

20%

17.87%, California

15%

10%
8.37%, Texas

5.52%, New York

5%
3.93%, Washington

0%
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Figure 4.1.10

Chapter 4 Preview 12
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.1 Jobs

AI Hiring in the same period the job began, divided by the total
number of LinkedIn members in the corresponding
Our AI hiring data is based on a LinkedIn dataset of skills
location. This rate is then indexed to the average
and jobs that appear on their platform. The countries
month in 2016; for example, an index of 1.1 in December
included in the sample make at least 10 AI hires each
2021 points to a hiring rate that is 10% higher than the
month and have LinkedIn covering at least 40% of
average month in 2016. LinkedIn makes month-to-
their labor force. India is also included in the sample
month comparisons to account for any potential lags in
given their increasing significance in the AI landscape,
members updating their profiles. The index for a year is
although LinkedIn does not cover 40% of their labor
the number in December of that year.
force. Therefore, the insights drawn about India should
be interpreted with particular caution.
The relative AI hiring index measures the degree to which
the hiring of AI talent is changing, more specifically
Figure 4.1.11 highlights the 15 geographic areas that
whether the hiring of AI talent is growing faster than,
have the highest relative AI hiring index for 2022. The
equal to, or more slowly than overall hiring in a particular
AI hiring rate is calculated as the percentage of LinkedIn
geographic region. In 2022, Hong Kong posted the
members with AI skills on their profile or working in
greatest growth in AI hiring at 1.4, followed by Spain, Italy
AI-related occupations who added a new employer
and the United Kingdom, and the United Arab Emirates.
Relative AI Hiring Index by Geographic Area, 2022
Source: LinkedIn, 2022 | Chart: 2023 AI Index Report

Hong Kong 1.37

Spain 1.19

Italy 1.18

United Kingdom 1.18

United Arab Emirates 1.15

South Africa 1.13

New Zealand 1.06

Denmark 1.06

Belgium 1.05

Netherlands 1.03

South Korea 1.02

Sweden 1.01

Canada 0.99

Switzerland 0.99

Singapore 0.99

0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40

Relative AI Hiring Index Figure 4.1.11

Figure 4.1.12 highlights how the AI hiring index changes over time for a wide range of countries3. Overall, the
majority of countries included in the sample have seen meaningful increases in their AI hiring rates since 2016.
This trend suggests that those countries are now hiring more AI talent than in 2016. However, for many countries,
AI hiring rates seem to have peaked around 2020, then dropped, and have since stabilized.
3 Both Figure 4.1.11 and Figure 4.1.12 report the Relative AI Hiring Index. Figure 4.1.11 reports the Index value at the end of December 2022, while Figure 4.1.12 reports a twelve-month rolling average.

Chapter 4 Preview 13
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.1 Jobs

Relative AI Hiring Index by Geographic Area, 2016–22

Source: LinkedIn, 2022 | Chart: 2023 AI Index Report

Australia Belgium Brazil Canada

2.00 2.00 2.00 2.00
1.50 1.50 1.50 1.50
1.12 1.11 1.17
1.00 1.00 1.00 1.00 1.00
0.50 0.50 0.50 0.50
0.00 0.00 0.00 0.00
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022

Chile Denmark Finland France

2.00 2.00 2.00 2.00
1.50 1.50 1.50 1.50
1.14 1.08 1.19
1.00 1.05 1.00 1.00 1.00
0.50 0.50 0.50 0.50
0.00 0.00 0.00 0.00
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022

Germany Hong Kong India Ireland

Relative AI Hiring Index

2.00 2.00 2.00 2.00

1.50 1.50 1.50 1.50
1.13 1.21 1.08
1.00 1.00 1.00 0.94 1.00
0.50 0.50 0.50 0.50
0.00 0.00 0.00 0.00
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022

Israel Italy Luxembourg Netherlands

2.00 2.00 2.00 2.00
1.50 1.50 1.50 1.50
1.11 1.21
1.00 1.00 1.00 1.00 1.03 1.00
0.50 0.50 0.50 0.50
0.00 0.00 0.00 0.00
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022

New Zealand Norway Portugal Singapore

2.00 2.00 2.00 2.00
1.50 1.50 1.50 1.50
1.09 1.10 1.01 1.13
1.00 1.00 1.00 1.00
0.50 0.50 0.50 0.50
0.00 0.00 0.00 0.00
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022

South Africa South Korea Spain Sweden

2.00 2.00 2.00 2.00
1.50 1.50 1.50 1.50
1.10 1.05 1.12 1.09
1.00 1.00 1.00 1.00
0.50 0.50 0.50 0.50
0.00 0.00 0.00 0.00
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022

Switzerland United Arab Emirates United Kingdom United States

2.00 2.00 2.00 2.00
1.50 1.50 1.50 1.50
1.25 1.15
1.00 1.09 1.00 1.08 1.00 1.00
0.50 0.50 0.50 0.50
0.00 0.00 0.00 0.00
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022
Figure 4.1.12

Chapter 4 Preview 14
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.1 Jobs

rate of various countries or regions from 2015 to

AI Skill Penetration 2022. In this case, the relative AI skill penetration rate
The AI skill penetration rate is a metric created by
can be understood as the sum of the penetration of
LinkedIn that measures the prevalence of various
each AI skill across occupations in a given country or
AI-related skills across occupations. LinkedIn
region, divided by the global average across the same
generates this metric by calculating the frequencies
occupation. For instance, a relative skill penetration
of LinkedIn users’ self-added skills in a given area
rate of 1.5 means that the average penetration of AI
from 2015 to 2022, then reweighting those numbers
skills in that country or region is 1.5 times the global
with a statistical model to create the top 50
average across the same set of occupations.
representative skills in that select occupation.
As of 2022, the three countries or regions with the
Global Comparison: Aggregate highest AI skill penetration rates were India (3.2),
Figure 4.1.13 shows the relative AI skill penetration the United States (2.2), and Germany (1.7).

Relative AI Skill Penetration Rate by Geographic Area, 2015–22

Source: LinkedIn, 2022 | Chart: 2023 AI Index Report

India 3.23

United States 2.23

Germany 1.72

Israel 1.65

Canada 1.54

United Kingdom 1.54

South Korea 1.44

Singapore 1.37

France 1.13

Brazil 0.99

Spain 0.98

Netherlands 0.95

Italy 0.95

Switzerland 0.91

Australia 0.89

0.00 0.50 1.00 1.50 2.00 2.50 3.00

Relative AI Skill Penetration Rate

Figure 4.1.13

Chapter 4 Preview 15
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.1 Jobs

Global Comparison: By Gender

Figure 4.1.14 disaggregates AI skill penetration rates pooled together across the same set of occupations
by gender across different countries or regions. in the country. For all countries in the sample, the
A country’s “Relative AI skill penetration rate relative AI skill penetration rate is greater for men
across genders” for women of 1.5 means that female than women. India (2.0), the United States (1.3), and
members in that country are 1.5 times more likely to Israel (0.9) have the highest reported relative AI skill
list AI skills than the average member in all countries penetration rates for women.

Relative AI Skill Penetration Rate Across Gender, 2015–22

Source: LinkedIn, 2022 | Chart: 2023 AI Index Report
3.27
India 1.99
2.36
United States 1.28
2.05
Israel 0.87
1.59
Canada 0.86
1.91
Germany 0.71
1.37
Singapore 0.68
1.46
United Kingdom 0.57
1.13
France 0.39
0.98
Netherlands 0.38
0.82
United Arab Emirates 0.31
0.85
Italy 0.31
0.98
Brazil 0.30
1.03
Spain 0.30
1.08 Male
Finland 0.29
Female
0.88
Australia 0.29

0.00 0.50 1.00 1.50 2.00 2.50 3.00

Relative AI Skill Penetration Rate
Figure 4.1.14

Chapter 4 Preview 16
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.2 Investment

Using data from NetBase Quid, this section tracks trends in AI-related investments. NetBase Quid tracks data on the investments of over
8 million global public and private companies. NetBase Quid also uses natural language processing techniques to search, analyze, and
identify patterns in large, unstructured datasets, like aggregated news and blogs, and company and patent databases. NetBase Quid
continuously broadens the set of companies for which it tracks data, so that in this year’s AI Index, the reported investment volume for
certain years is larger than that of previous reports.

4.2 Investment
Corporate Investment
As AI becomes more and more integrated into the For the first time since 2013, year-over-year global
economy, it becomes increasingly important to track corporate investment in AI has decreased. In 2022,
AI-related corporate investment. Figure 4.2.1 shows total global corporate AI investment was $189.6
overall global corporate investment in AI from 2013 billion, roughly a third lower than it was in 2021.
to 2022. Corporate investment includes mergers and Still, in the last decade, AI-related investment has
acquisitions, minority stakes, private investment, and increased thirteenfold.
public offerings.

Global Corporate Investment in AI by Investment Activity, 2013–22

Source: NetBase Quid, 2022 | Chart: 2023 AI Index Report
300
Merger/Acquisition
Minority Stake 276.14
Private Investment
250 Public O ering
Total Investment (in Billions of U.S. Dollars)

119.66
200
189.59

150 146.74 83.35

26.06

100 95.57 46.06

125.36
79.62
31.91
21.89
53.72
50 91.86
33.82 24.68 61.61
25.43 43.1 55.09
19.04 13.05
14.57
17.13 25.72 29.1
13.35 12.62 13.01
0
2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Figure 4.2.1

Chapter 4 Preview 17
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.2 Investment

To provide a fuller context for the Top Five AI Merger/Acquisition Investment Activities, 2022
Source: NetBase Quid, 2022 | Table: 2023 AI Index Report
nature of AI investment in the last year,
Company Name Headquarters Focus Area Funding Amount
Figures 4.2.2 through 4.2.5 highlight Country (in Billions USD)

the top merger/acquisition, minority Nuance United States Arti cial Intelligence; 19.80
Communications, Inc. Enterprise Software;
stake, private investment, and public Healthcare; Machine
offering events in the last year. The Learning

greatest single AI investment event Citrix Systems, Inc. United States Data Management, 17.18
Processing, and Cloud;
was the merger/acquisition of Nuance HR Tech

Communications, valued at $19.8 billion Avast Limited Czech Republic Data Management, 8.02
Processing, and Cloud;
(Figure 4.2.2). The largest minority Fintech; Cybersecurity,
stake event was for the British company Data Protection

6.34
Aveva Group ($4.7 billion) (Figure 4.2.3). AspenTech United States Manufacturing;
Corporation Software; Supply
The greatest private investment event Chain Management

was GAC Aion New Energy Automobile Vivint Smart Home, United States Cybersecurity, Data 5.54
Inc. Protection; Sales
($2.5 billion), a Chinese clean energy Enablement
and automotive company (Figure 4.2.4).
Finally, the largest public offering was Figure 4.2.2

ASR Microelectronics ($1.1 billion),

a Chinese semiconductor company Top Five AI Minority Stake Investment Activities, 2022
Source: NetBase Quid, 2022 | Table: 2023 AI Index Report
(Figure 4.2.5).
Company Name Headquarters Focus Area Funding Amount
Country (in Billions USD)

AVEVA Group, PLC United Chemical; Computer; 4.68

Kingdom Data Mining; Electronics;
Industrial Manufacturing;
Information Technology;
Simulation; Software

Grupo de Colombia Financial Services; Impact 1.48

Inversiones Investing; Insurance
Suramericana, SA

Fractal Analytics India Analytics; Arti cial 0.35

Private Limited Intelligence; Big Data;
Business Intelligence;
Consulting; Machine
Learning

Atrys Health, SA Spain Medical and Healthcare 0.28

R Systems India Analytics; Information 0.17

International, Ltd. Technology; IT
Management; Software

Figure 4.2.3

Chapter 4 Preview 18
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.2 Investment

Top Five AI Private Investment Activities, 2022

Source: NetBase Quid, 2022 | Table: 2023 AI Index Report

Company Name Headquarters Focus Area Funding Amount

Country (in Billions USD)

GAC Ai´an New China Automotive; Clean 2.54

Energy Automobile Energy; Electric Vehicle;
Co., Ltd. Manufacturing

Idience Co., Ltd. South Korea Emergency Medicine; 2.15

Healthcare;
Pharmaceutical

Uali Argentina Drones; Cloud Computing 1.50

Anduril Industries, United States Cybersecurity, Data 1.50

Inc. Protection; AR/VR;
Drones

Celonis, GmbH Germany Retail; Industrial 1.22

Automation, Network; HR
Tech; Insurtech

Figure 4.2.4

Top Five AI Public O ering Investment Activities, 2022

Source: NetBase Quid, 2022 | Table: 2023 AI Index Report

Company Name Headquarters Focus Area Funding Amount

Country (in Billions USD)

ASR Microelectronics China Semiconductor; VC 1.08

Co., Ltd.

iSoftStone Information China Data Management, 0.73

Technology (Group) Processing, and Cloud;
Co., Ltd. Cybersecurity, Data
Protection

Jahez International Saudi Arabia Arti cial Intelligence; 0.43

Company for E-Commerce; Food and
Information Systems Beverage; Food Delivery;
Technology Information Technology;
Logistics

Fortior Technology China Electronics; Machine 0.30

(Shenzhen) Co., Ltd. Manufacturing;
Semiconductor

Beijing Deep Glint China Cybersecurity, Data 0.29

Technology Co., Ltd. Protection; Music, Video
Content

Figure 4.2.5

Chapter 4 Preview 19
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.2 Investment

Startup Activity
Global Trend
The next section analyzes private investment trends in The global private AI investment trend reveals that
artificial intelligence startups that have received over while investment activity has decreased since 2021, it
$1.5 million in investment since 2013. is still 18 times higher than it was in 2013 (Figure 4.2.6).

Private Investment in AI, 2013–22

Source: NetBase Quid, 2022 | Chart: 2023 AI Index Report

120
Total Investment (in Billions of U.S. Dollars)

100
91.86

80

60

40

20

0
2013 2014 2015 2016 2017 2018 2019 2020 2021 2022

Figure 4.2.6

Chapter 4 Preview 20
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.2 Investment

A similar trend, of short-term decreases but longer- decrease from 2021 but a sixfold increase since 2013
term growth, is evident in data on total private (Figure 4.2.7). Similarly, the number of newly funded AI
investment events. In 2022 there were 3,538 AI- companies dropped to 1,392 from 1,669 last year, while
related private investment events, representing a 12% having increased from 495 in 2013 (Figure 4.2.8).
Number of Private Investment Events in AI, 2013–22
Source: NetBase Quid, 2022 | Chart: 2023 AI Index Report

4,000

3,538
3,500
Number of Private Investment Events

3,000

2,500

2,000

1,500

1,000

500

0
2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Figure 4.2.7
Number of Newly Funded AI Companies in the World, 2013–22
Source: NetBase Quid, 2022 | Chart: 2023 AI Index Report

1,600

1,392
1,400

1,200
Number of Companies

1,000

800

600

400

200

0
2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Figure 4.2.8

Chapter 4 Preview 21
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.2 Investment

The year-over-year decrease in AI-related AI Private Investment Events by Funding Size,

2021 Vs. 2022
funding is also evident when the funding events Source: NetBase Quid, 2022 | Table: 2023 AI Index Report

are disaggregated by size. Across all size Funding Size 2021 2022 Total

categories, with the exception of ones over Over $1 Billion 4 6 10

$500 Million – $1 Billion 13 5 18
$1 billion, the total number of AI funding events
$100 Million – $500 Million 277 164 441
decreased (Figure 4.2.9). $50 Million – $100 Million 277 238 515
Under $50 Million 2,851 2,585 5,436
Undisclosed 598 540 1,138
Total 4,020 3,538 7,558

Figure 4.2.9

Regional Comparison by Funding Amount

Once again, the United States led the world in terms of total AI private investment. In 2022, the $47.4 billion
invested in the United States was roughly 3.5 times the amount invested in the next highest country, China
($13.4 billion), and 11 times the amount invested in the United Kingdom ($4.4 billion) (Figure 4.2.10).

Private Investment in AI by Geographic Area, 2022

Source: NetBase Quid, 2022 | Chart: 2023 AI Index Report

United States 47.36

China 13.41

United Kingdom 4.37

Israel 3.24

India 3.24

South Korea 3.10

Germany 2.35

Canada 1.83

France 1.77

Argentina 1.52

Australia 1.35

Singapore 1.13

Switzerland 1.04

Japan 0.72

Finland 0.61

0 5 10 15 20 25 30 35 40 45
Total Investment (in Billions of U.S. Dollars)
Figure 4.2.10

Chapter 4 Preview 22
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.2 Investment

When private AI investments are aggregated since 2013, the same ranking of countries applies:
The United States is first with $248.9 billion invested, followed by China ($95.1 billion) and the
United Kingdom ($18.2 billion) (Figure 4.2.11).

Private Investment in AI by Geographic Area, 2013–22 (Sum)

Source: NetBase Quid, 2022 | Chart: 2023 AI Index Report

United States 248.90

China 95.11

United Kingdom 18.24

Israel 10.83

Canada 8.83

India 7.73

Germany 6.99

France 6.59

South Korea 5.57

Singapore 4.72

Japan 3.99

Hong Kong 3.10

Switzerland 3.04

Australia 3.04

Spain 1.81

0 20 40 60 80 100 120 140 160 180 200 220 240

Total Investment (in Billions of U.S. Dollars)
Figure 4.2.11

Chapter 4 Preview 23
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.2 Investment

While the United States continues to outpace The top five American AI private investment events
other nations in terms of private AI investment, are highlighted in Figure 4.2.13, the top five European
the country experienced a sharp 35.5% decrease Union and British investments in Figure 4.2.14, and the
in AI private investment within the last year (Figure top five Chinese investments in Figure 4.2.15.
4.2.12). Chinese investment experienced a similarly
sharp decline (41.3%).

Private Investment in AI by Geographic Area, 2013–22

Source: NetBase Quid, 2022 | Chart: 2023 AI Index Report

70

60
Total Investment (in Billions of U.S. Dollars)

50
47.36, United States

40

30

20

13.41, China
10 11.04, European Union and United Kingdom

2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Figure 4.2.12

Chapter 4 Preview 24
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.2 Investment

Top AI Private Investment Events in the United Top AI Private Investment Events in the European
States, 2022 Union and United Kingdom, 2022
Source: NetBase Quid, 2022 | Table: 2023 AI Index Report Source: NetBase Quid, 2022 | Table: 2023 AI Index Report

Company Name Focus Area Funding Amount Company Name Focus Area Funding Amount
(in Billions USD) (in Billions USD)

Anduril Industries, Inc. Cybersecurity, Data 1.50 Celonis, GmbH Retail; Industrial 1.22
Protection; AR/VR; Automation, Network;
Drones HR Tech; Insurtech

Faire Wholesale, Inc. Fintech; Retail; Sales 0.82 Content Square, SAS Analytics; Arti cial 0.60
Enablement Intelligence: CRM:
Data Visualization;
Anthropic, PBC Arti cial Intelligence; 0.58 Digital Marketing;
Information SaaS
Technology; Machine
Learning Retail Logistics Excellence Retail 0.57
- RELEX Oy
Arctic Wolf Networks, Inc. Data Management, 0.40
Processing, and Cloud; Cera Care Limited Medical and 0.32
Cybersecurity, Data Healthcare
Protection
Babylon Holdings Limited Medical and 0.30
JingChi, Inc. Data Management, 0.40 Healthcare; Music,
Processing, and Cloud; Video Content
AV; AR/VR
Figure 4.2.14
Figure 4.2.13

Top AI Private Investment Events in China, 2022

Source: NetBase Quid, 2022 | Table: 2023 AI Index Report

Company Name Focus Area Funding Amount

(in Billions USD)

GAC Ai´an New Energy Automotive; Clean 2.54

Automobile Co., Ltd. Energy; Electric
Vehicle;
Manufacturing

GAC Ai´an New Energy Automotive; Clean 1.11

Automobile Co., Ltd. Energy; Electric
Vehicle;
Manufacturing

Beijing ESWIN Data Management, 0.58

Technology Group Co., Processing, and Cloud;
Ltd. Industrial Automation,
Network;
Semiconductor;
Marketing, Digital Ads;
Sales Enablement

Zhejiang Hozon New Data Management, 0.44

Energy Automobile Co., Processing, and Cloud;
Ltd. Cybersecurity, Data
Protection; Sales
Enablement

Zhejiang Hozon New Data Management, 0.32

Energy Automobile Co., Processing, and Cloud;
Ltd. Cybersecurity, Data
Protection; Sales
Enablement

Figure 4.2.15

Chapter 4 Preview 25
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.2 Investment

Regional Comparison by Newly Funded

AI Companies
This subsection studies the number of newly funded United States led all regions with the largest number of
AI companies across various geographic areas. newly funded AI companies at 542, followed by China
As was the case with private investment, the at 160 and the United Kingdom at 99 (Figure 4.2.16).

Number of Newly Funded AI Companies by Geographic Area, 2022

Source: NetBase Quid, 2022 | Chart: 2023 AI Index Report

United States 542

China 160

United Kingdom 99

Israel 73

India 57

Canada 47

France 44

Germany 41

Singapore 36

Japan 32

Switzerland 26

Australia 23

South Korea 22

Sweden 12

Netherlands 12

0 100 200 300 400 500

Number of Companies
Figure 4.2.16

Chapter 4 Preview 26
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.2 Investment

A similar trend is evident in the aggregate data since 2013. In the last decade, the number of newly funded
AI companies in the United States is around 3.5 times the amount in China, and 7.4 times the amount in the
United Kingdom (Figure 4.2.17).

Number of Newly Funded AI Companies by Geographic Area, 2013–22 (Sum)

Source: NetBase Quid, 2022 | Chart: 2023 AI Index Report

United States 4,643

China 1,337

United Kingdom 630

Israel 402

Canada 341

France 338

India 296

Japan 294

Germany 245

Singapore 165

South Korea 145

Australia 126

Switzerland 108

Sweden 83

Netherlands 78

0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500

Number of Companies Figure 4.2.17

Figure 4.2.18 breaks Number of Newly Funded AI Companies by Geographic Area, 2013–22
Source: NetBase Quid, 2022 | Chart: 2023 AI Index Report
700
down data on newly
funded AI companies
within select 600

geographic regions. 542,

United States
In a trend that goes 500

back a decade,
Number of Companies

the United States 400

continues to outpace
both the European 300 293,
European
Union and the United Union and
United Kingdom
Kingdom, as well as 200
160,
China. However, the China
growth rates of the 100

different regions are

relatively similar. 0
2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Figure 4.2.18

Chapter 4 Preview 27
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.2 Investment

Focus Area Analysis Private Investment in AI by Focus Area, 2021 Vs. 2022
Source: NetBase Quid, 2022 | Chart: 2023 AI Index Report
Private AI investment can also be disaggregated by Medical and Healthcare
Data Management, Processing, Cloud
focus area. Figure 4.2.19 compares global private Fintech
AI investment by focus area in 2022 versus 2021. Cybersecurity, Data Protection
Retail
The focus areas that attracted the most investment Industrial Automation, Network
Sales Enablement
in 2022 were medical and healthcare ($6.1 billion); Marketing, Digital Ads
AR/VR
data management, processing, and cloud ($5.9 Drones
Insurtech
billion); fintech ($5.5 billion); cybersecurity and Music, Video Content
data protection ($5.4 billion); and retail ($4.2 Semiconductor
HR Tech
billion). Mirroring the pattern seen in total AI private Energy, Oil, and Gas
AV
investment, the total investment across most focus NLP, Customer Support
Agritech
areas declined in the last year. Entertainment
Legal Tech
Geospatial
Fitness and Wellness
Ed Tech 2022
Facial Recognition 2021
VC
0 2 4 6 8 10
Total Investment (in Billions of U.S. Dollars)
Figure 4.2.19

Figure 4.2.20 presents trends in AI focus area cybersecurity and data protection, drones, marketing
investments. As noted earlier, most focus areas saw and digital ads, HR tech, AR/VR, and legal tech. Still,
declining investments in the last year. However, some mirroring a broader trend in AI private investment,
of the focus areas that saw increased investments are most focus areas saw greater amounts of AI private
semiconductor, industrial automation and network, investment in 2022 than they did in 2017.

Chapter 4 Preview 28
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.2 Investment

Private Investment in AI by Focus Area, 2017–22

Source: NetBase Quid, 2022 | Chart: 2023 AI Index Report

Data Management, Processing, Cloud Medical and Healthcare Fintech AV

8 8 8 8
6 5.86 6 6.05 6 5.52 6
4 4 4 4
2 2 2 2 1.34
0 0 0 0
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022

Semiconductor Industrial Automation, Network Retail Fitness and Wellness

8 8 8 8
6 6 6 6
4 4 3.92 4 4.20 4
2 1.65 2 2 2
0.53
0 0 0 0
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022

NLP, Customer Support Energy, Oil, and Gas Cybersecurity, Data Protection Drones

8 8 8 8
6 6 6 5.38 6
4 4 4 4
2 2 1.61 2 2 1.88
1.01
Total Investment (in Billions of U.S. Dollars)

0 0 0 0
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022

Marketing, Digital Ads HR Tech Facial Recognition Insurtech

8 8 8 8
6 6 6 6
4 4 4 4
3.05
2 2 1.63 2 2 1.74
0.07
0 0 0 0
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022

Agritech Sales Enablement AR/VR Ed Tech

8 8 8 8
6 6 6 6
4 4 4 4
3.18 2.39
2 2 2 2
0.87 0.37
0 0 0 0
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022

Geospatial Legal Tech Entertainment Music, Video Content

8 8 8 8
6 6 6 6
4 4 4 4
2 2 2 2 1.72
0.71 0.83 0.87
0 0 0 0
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022

VC

8
6
4
2
0.02
0 Figure 4.2.20
2018 2020 2022

Chapter 4 Preview 29
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.2 Investment

Finally, 4.2.21 shows private investment in AI by focus area billion), and 40 times more than that in the European
over time within select geographic regions, highlighting Union and the United Kingdom ($0.04 billion). Chinese
how private investment priorities in AI differ across private investment in AI-related semiconductors ($1.02
geographies. For example, in 2022, private investment billion) was 1.75 times more than that in the United
in AI-related drone technology in the United States ($1.6 States ($0.58 billion), and 102 times more than that in the
billion) was nearly 53 times more than that in China ($0.03 European Union and the United Kingdom ($0.01 billion).

Private Investment in AI by Focus Area and Geographic Area, 2017–22

Source: NetBase Quid, 2022 | Chart: 2023 AI Index Report

Data Management, Processing, Cloud Medical and Healthcare Fintech AV

US, 3.13 US, 4.19 US, 3.23 US, 0.69
CN, 1.87 CN, 0.25 CN, 0.03 CN, 0.49
4 EU/UK, 0.24 4 EU/UK, 0.76 4 EU/UK, 0.94 4 EU/UK, 0.02
2 2 2 2

0 0 0 0
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022

Semiconductor Industrial Automation, Network Retail Fitness and Wellness

US, 0.58 US, 0.87 US, 1.52 US, 0.23
CN, 1.02 CN, 1.06 CN, 0.01 CN, 0.00
4 EU/UK, 0.01 4 EU/UK, 1.65 4 EU/UK, 2.07 4 EU/UK, 0.14
2 2 2 2

0 0 0 0
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022

NLP, Customer Support Energy, Oil, and Gas Cybersecurity, Data Protection Drones
US, 0.69 US, 0.80 US, 3.87 US, 1.60
CN, 0.13 CN, 0.34 CN, 1.07 CN, 0.03
4 EU/UK, 0.04 4 EU/UK, 0.20 4 EU/UK, 0.23 4 EU/UK, 0.04
2 2 2 2

0 0 0 0
Total Investment (in Billions of U.S. Dollars)

2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022
Marketing, Digital Ads HR Tech Facial Recognition Insurtech
US, 1.14 US, 0.24 US, 0.07 US, 0.39
CN, 0.88 CN, 0.00 CN, 0.00 CN, 0.00
4 EU/UK, 0.76 4 EU/UK, 1.28 4 EU/UK, 0.00 4 EU/UK, 1.29
2 2 2 2

0 0 0 0
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022

Agritech Sales Enablement AR/VR Ed Tech

US, 0.55 US, 1.12 US, 2.07 US, 0.12
CN, 0.10 CN, 1.68 CN, 0.01 CN, 0.01
4 EU/UK, 0.08 4 EU/UK, 0.16 4 EU/UK, 0.06 4 EU/UK, 0.10
2 2 2 2

0 0 0 0
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022

Geospatial Legal Tech Entertainment Music, Video Content

US, 0.55 US, 0.71 US, 0.47 US, 1.10
CN, 0.03 CN, 0.05 CN, 0.18 CN, 0.03
4 EU/UK, 0.01 4 EU/UK, 0.06 4 EU/UK, 0.17 4 EU/UK, 0.44
2 2 2 2

0 0 0 0
2018 2020 2022 2018 2020 2022 2018 2020 2022 2018 2020 2022

VC
US, 0.00
CN, 0.00
4 EU/UK, 0.02
2

0 Figure 4.2.21
2018 2020 2022

Chapter 4 Preview 30
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

This section explores how corporations tangibly use AI. First, it highlights industry adoption trends and asks how businesses adopt
AI and what particular AI technologies they find most useful, and identifies how AI adoption affects their bottom line. Second, the
section considers industry motivations and explores what questions industry leaders consider when thinking about incorporating AI
technologies. Finally, it paints a qualitative picture of business AI use by examining trends in AI-related earnings calls.

4.3 Corporate Activity

Industry Adoption
Adoption of AI Capabilities
The following subsection on the industry adoption According to the most recent McKinsey report, as of
of AI borrows data from McKinsey’s “The State of 2022, 50% of surveyed organizations reported having
AI in 2022—and a Half Decade in Review,” as well adopted AI in at least one business unit or function
as previous years’ editions. The 2022 report drew on (Figure 4.3.1). This total is down slightly from 56% in
data from a survey of 1,492 participants representing 2021, although up significantly from 20% in 2017. AI
a wide range of regions, industries, company sizes, usage has rapidly grown in the past half-decade, but
functional specialties, and tenures. leveled off since 2020.

Share of Respondents Who Say Their Organizations Have Adopted AI in at Least One Function, 2017–22
Source: McKinsey & Company Survey, 2022 | Chart: 2023 AI Index Report
60%

50% 50%

40%
% of Respondents

30%

20%

10%

0%
2017 2018 2019 2020 2021 2022

Figure 4.3.1

Chapter 4 Preview 31
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

In the last half-decade, the average number of AI capabilities that organizations have embedded
has doubled from 1.9 in 2018 to 3.8 in 2022 (Figure 4.3.2). Some of the AI capabilities that McKinsey
features in their survey include recommender systems, NL text understanding, and facial recognition.4

Average Number of AI Capabilities That Respondents’ Organizations Have Embedded Within at Least One
Function or Business Unit, 2018–22
Source: McKinsey & Company Survey, 2022 | Chart: 2023 AI Index Report

4.00
3.80

3.50
Number of AI Capabilities (Average)

3.00

2.50

2.00

1.50

1.00

0.50

0.00
2018 2019 2020 2021 2022

Figure 4.3.2

4 In the 2022 edition of the McKinsey survey, 16 total AI capabilities are considered: computer vision, deep learning, digital twins, facial recognition, GAN, knowledge graphs,
NL generation, NL speech understanding, NL text understanding, physical robotics, recommender systems, reinforcement learning, robotic process automation, transfer
learning, transformers, and virtual agents.

Chapter 4 Preview 32
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

The most commonly adopted AI use case in 2022 was service operations optimization (24%), followed
by the creation of new AI-based products (20%), customer segmentation (19%), customer service
analytics (19%), and new AI-based enhancement of products (19%) (Figure 4.3.3).

Most Commonly Adopted AI Use Cases by Function, 2022

Source: McKinsey & Company Survey, 2022 | Chart: 2023 AI Index Report

Service Operations Optimization 24%

Creation of New AI-Based Products 20%

Customer Segmentation 19%

Customer Service Analytics 19%

New AI-Based Enhancements of Products 19%

Customer Acquisition and Lead Generation 17%

Product Feature Optimization 16%

Contact-Center Automation 16% Service Operations

Product and/or
Risk Modeling and Analytics 15% Service Development
Marketing and Sales
Predictive Service and Intervention 14% Risk

0% 4% 8% 12% 16% 20% 24%

% of Respondents

Figure 4.3.3

Chapter 4 Preview 33
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

With respect to the type of AI capabilities embedded rates of embedding were 48%, 47%, and 46%. Across
in at least one function or business unit, as indicated all industries, the most embedded AI technologies
by Figure 4.3.4, robotic process automation had were robotic process automation (39%), computer
the highest rate of embedding within high tech/ vision (34%), NL text understanding (33%), and virtual
telecom, financial services and business, and legal agents (33%).
and professional services industries—the respective

AI Capabilities Embedded in at Least One Function or Business Unit, 2022

Source: McKinsey & Company Survey, 2022 | Chart: 2023 AI Index Report

All Industries 34% 30% 24% 18% 11% 25% 18% 23% 33% 20% 25% 20% 39% 16% 11% 33%

Business, Legal, and

32% 37% 31% 11% 8% 26% 12% 22% 34% 19% 23% 26% 46% 16% 11% 30%
Professional Services

Consumer Goods/
33% 36% 25% 19% 13% 18% 20% 11% 22% 24% 32% 19% 25% 7% 11% 40%
Industry

Retail

Financial Services 24% 22% 18% 24% 13% 29% 20% 30% 42% 14% 30% 19% 47% 17% 12% 33%

Healthcare Systems/
Pharma and 32% 18% 16% 5% 5% 14% 5% 12% 29% 11% 16% 13% 16% 9% 6% 14%
Med. Products

High Tech/Telecom 37% 45% 24% 16% 15% 23% 24% 29% 40% 15% 34% 23% 48% 22% 15% 43%
C

Fa

Kn

Ph

Re

Re

Ro

Tr

Tr

Vi
ee

ig
om

AN

an

an

rt
c

co

i
ys

bo
ow

nf
ita

Sp

Te
ia

ua
p

sf

sf
ic
pu

or
en

tic
lR

le

xt
Le

lT

lA
er

or
al

ce
ec
te

dg

er
ec

Un

m
w

Pr
a

Ro

Le

ge
en
r

rn

m
h
at
in

og

er
oc
Vi

ar
de

nt
b

en
Un
in

io
s

s(
er
ni

ot
sio

es

ni

s
g

n
ra

rS

tL
d

st

e.
tio

ic

ng
sA
er
ph
n

an

g.
ys
s

ea
n

st

,G
ut
s

di

te

rn
an

om
ng

in

PT
di

at

-3
ng

io

)
n

% of Respondents (AI Capability)

Figure 4.3.4

Chapter 4 Preview 34
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

Figure 4.3.5 shows AI adoption by industry and AI function in 2022. The greatest adoption was in risk for high
tech/telecom (38%), followed by service operations for consumer goods/retail (31%) and product and/or service
development for financial services (31%).

AI Adoption by Industry and Function, 2022

Source: McKinsey & Company Survey, 2022 | Chart: 2023 AI Index Report

All Industries 11% 8% 5% 10% 19% 19% 21% 9%

Business, Legal, and

11% 10% 9% 8% 16% 20% 19% 12%
Professional Services

Consumer Goods/
14% 4% 3% 4% 15% 31% 29% 11%
Retail
Industry

Financial Services 1% 8% 7% 31% 17% 24% 23% 2%

Healthcare Systems/
Pharma and 15% 7% 2% 4% 22% 12% 8% 8%
Med. Products

High Tech/Telecom 6% 6% 4% 7% 38% 21% 25% 8%

Hu Ma Ma Pro Ris Se Str Su

rv p
ma
nR
nu
fa
rke
ti Se duct k ice Co ateg Ma ply-
eso ctu ng rvi
c an Op rpo y a na Ch
urc rin
g
an
dS e D d/o era rat nd ge a
me in
es ev r tio eF nt
ale elo ns ina
s pm nce
en
t
% of Respondents (Function)
Figure 4.3.5

Chapter 4 Preview 35
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

Figure 4.3.6 shows how rates of AI adoption by points); followed by high tech/telecom, for risk
industry and AI function vary from 2021 to 2022 (22 percentage points). The most significant
in order to demonstrate how rates of AI adoption decreases were in high tech/telecom, for product
have changed over the last year. The greatest year- and/or service development (38 percentage points);
over-year increases were in consumer goods/retail, and healthcare systems, also for product and/or
for strategy and corporate finance (25 percentage service development (25 percentage points).

Percentage Point Change in Responses of AI Adoption by Industry and Function 2021 Vs. 2022
Source: McKinsey & Company Survey, 2022 | Chart: 2023 AI Index Report

All Industries 2% -4% -15% -13% 6% -6% 12% -4%

Business, Legal, and

-3% 2% -19% -7% 3% -6% 11% -1%
Professional Services

Consumer Goods/
12% -14% -19% -13% 14% 16% 25% -7%
Retail
Industry

Financial Services -9% 4% -17% 11% -15% -16% 10% -6%

Healthcare Systems/
Pharma and 6% -4% -12% -25% 9% -5% -4% -1%
Med. Products

High Tech/Telecom -6% -5% -24% -38% 22% -13% 15% -8%

Hu Ma Ma Pro Ris Se Str Su

rv p
ma
nR
nu
fa
rke
ti Se duct k ice Co ateg Ma ply-
eso ctu ng rvi
ce a nd Op rpo y a na Ch
urc rin
g
an
dS De /or era rat nd ge a
me in
es ve tio eF nt
ale lop ns ina
s me nce
nt
Percentage Point Change in Responses (Function)
Figure 4.3.6

Chapter 4 Preview 36
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

Organizations report AI adoption leading to both (Figure 4.3.7). On the revenue side, the functions that
cost decreases and revenue increases. On the cost most respondents saw increases in as a result of AI
side, the functions that most respondents saw adoption were marketing and sales (70%), product
decreases in as a result of AI adoption were supply and/or service development (70%), and strategy and
chain management (52%), service operations (45%), corporate finance (65%).
strategy and corporate finance (43%), and risk (43%)

Cost Decrease and Revenue Increase From AI Adoption by Function, 2021

Source: McKinsey & Company Survey, 2022 | Chart: 2023 AI Index Report

Decrease by <10% Decrease by 10–19% Decrease by ≥20% Increase by >10% Increase by 6–10% Increase by ≤5%

Service Operations 45% 29% 10% 6% 10% 10% 37% 57%

Manufacturing 42% 32% 7% 10% 18% 33% 61%

Human Resources 29% 25% 14% 13% 31% 58%

Marketing and Sales 28% 21% 9% 20% 41% 70%

Function

Risk 43% 30% 8% 10% 11% 27% 48%

Supply Chain Management 52% 41% 7% 14% 17% 28% 59%

Product and/or Service Development 30% 20% 6% 13% 24% 33% 70%

Strategy and Corporate Finance 43% 31% 8% 8% 16% 41% 65%

Average Across All Activities 32% 23% 8% 19% 36% 63%

% of Respondents

Figure 4.3.7

Chapter 4 Preview 37
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

Figure 4.3.8 shows AI adoption by organizations across all geographies was 50%, down 6% from 2021.
globally, broken out by regions of the world. In 2022, Notably, “Greater China” registered a 20 percentage
North America led (59%), followed by Asia-Pacific point decrease from 2021.
(55%) and Europe (48%). The average adoption rate

AI Adoption by Organizations in the World, 2021 Vs. 2022

Source: McKinsey & Company Survey, 2022 | Chart: 2023 AI Index Report

50%
All Geographies
56%

55%
Asia-Paci c
64%

48%
Europe
51%

59%
North America
55%

Greater China 41%

(incl. Hong Kong,
Taiwan) 61%

Developing Markets 2022

(incl. India, 44%
2021
Latin America, 52%
MENA)
0% 10% 20% 30% 40% 50% 60%
% of Respondents
Figure 4.3.8

Chapter 4 Preview 38
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

Consideration and Mitigation of Risks From risks were regulatory compliance (45%), personal/
Adopting AI individual privacy (40%), and explainability (37%).
As has been the case in the last few iterations of the The least salient risks identified by organizations were
McKinsey report, in 2022 respondents identified national security (13%) and political stability (9%).
cybersecurity as the most relevant risk when adopting
AI technology (59%) (Figure 4.3.9). The next most cited

Risks From Adopting AI That Organizations Consider Relevant, 2019–22

Source: McKinsey & Company Survey, 2022 | Chart: 2023 AI Index Report

60% 59%, Cybersecurity

50%

45%, Regulatory Compliance

% of Respondents

40% 40%, Personal/Individual Privacy

37%, Explainability

32%, Organizational Reputation

30% 30%, Equity and Fairness
28%, Workforce/Labor Displacement

20% 20%, Physical Safety

13%, National Security

10%
9%, Political Stability

0%
2019 2020 2021 2022

Figure 4.3.9

Chapter 4 Preview 39
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

Figure 4.3.10 highlights the AI risks that organizations have taken steps to mitigate. For instance, there is
are taking steps to mitigate. The top three responses a gap of 8 percentage points for cybersecurity, 9
were cybersecurity (51%), followed by regulatory percentage points for regulatory compliance, and 12
compliance (36%) and personal/individual privacy percentage points for personal/individual privacy.
(28%). As was the case in previous years, there are These differences suggest there is a gap between
meaningful gaps between the risks organizations the awareness organizations have of various risks and
cite as relevant and those which organizations their steps taken to mitigate such risks.

Risks From Adopting AI That Organizations Take Steps to Mitigate, 2019–22

Source: McKinsey & Company Survey, 2022 | Chart: 2023 AI Index Report

51%, Cybersecurity
50%

40%

36%, Regulatory Compliance

% of Respondents

30%
28%, Personal/Individual Privacy

22%, Organizational Reputation

22%, Explainability
20%
18%, Workforce/Labor Displacement
17%, Equity and Fairness
15%, Physical Safety

10%
7%, National Security
4%, Political Stability

0%
2019 2020 2021 2022

Figure 4.3.10

Chapter 4 Preview 40
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

Narrative Highlight:
The Effects of GitHub’s Copilot on Developer
Productivity and Happiness
In 2021, launched a technical preview of Copilot,
It took the developers
a generative AI tool that enables developers and
coders to present a coding problem in natural using Copilot only 71
language and then have Copilot generate a minutes to complete their
solution in code. Copilot can also translate
between various programming languages. In
task—56% less time than
2022, GitHub surveyed over 2,000 developers the developers who did not
who were using the tool to determine its effect on
their productivity, well-being, and workflow.5
use Copilot (161 minutes).
Figure 4.3.11 summarizes the results of the survey. reported a completion rate of 78%, 8 percentage
Developers overwhelmingly reported feeling points higher than those who did not use Copilot.
more productive, satisfied, and efficient when Likewise, it only took the developers using Copilot
working with Copilot. More specifically, 88% of 71 minutes to complete their task, which was 56%
surveyed respondents commented feeling more less time than the developers who did not use
productive, 74% reported being able to focus on Copilot (161 minutes). These survey and experiment
more satisfying work, and 88% claimed to have results are evidence of the tangible ways in which
completed tasks more quickly. One software AI tools improve worker productivity.
engineer stated, “[With Copilot] I have to think
less, and when I have to think, it’s the fun stuff. It
sets off a little spark that makes coding more fun
and more efficient.”6

As part of the same survey, GitHub recruited

95 developers and randomly split them into two
groups, one of which used Copilot as part of a
coding task and the other which did not. The
results of this experiment are summarized in
Figure 4.3.12. The developers who used Copilot

5 Most of the developers surveyed, around 60%, were professional developers; 30% were students and 7% were hobbyists.
6 The quote is taken from this source.

Chapter 4 Preview 41
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

Narrative Highlight:
The Effects of GitHub’s Copilot on Developer
Productivity and Happiness (cont’d)

Measuring Dimensions of Developer Productivity When Using Copilot: Survey Responses, 2022
Source: GitHub Survey, 2022 | Chart: 2023 AI Index Report

I Am More Productive 88%

Focus on More Satisfying Work 74%

More Ful�lled With My Job 60% Perceived Productivity

Satisfaction and Well-Being
Less Frustrated When Coding 59% E�ciency and Flow

Faster With Repetitive Tasks 96%

Faster Completion 88%

Less Mental E�orts on Repetitive Tasks 87%

Less Time Searching 77%

More In The Flow 73%

0% 20% 40% 60% 80% 100%

% of Participants That Agreed or Strongly Agreed

Figure 4.3.11

Summary of the Experiment Process and Results

Source: GitHub Survey, 2022 | Table: 2023 AI Index Report

Used Did Not Use

GitHub Copilot GitHub Copilot

Number of Developers 45 50
Completion Rate (%) 78 70

Average Time Taken to 71 161

Complete the Task (Minutes)

Figure 4.3.12

Chapter 4 Preview 42
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

Importance of AI Solutions for Organizations’

Industry Motivation Overall Success
Source: Deloitte Survey, 2022 | Chart: 2023 AI Index Report
This section explores the motivations industry
leaders have in deploying AI and examines the
1%, Not Important
degree to which they feel AI is important, the 5%, Somewhat Important
reasons they are eager to embrace AI, and the
factors that have hindered further scaling of
AI solutions. The data from this section comes
from Deloitte’s “State of AI in Enterprise” report,
which has surveyed companies about their use
of AI since 2017. This year’s survey polled 2,620 94%, Important
business leaders from a wide range of countries,
industries, and corporate levels.
Figure 4.3.13

Perceived Importance of AI
Figures 4.3.13 and 4.3.14 suggest that an
overwhelming majority of business leaders
perceive AI to be important for their businesses.
More specifically, when asked how important
AI solutions were for their organization’s overall Believe AI Enhances Performance and Job
Satisfaction, 2022
Source: Deloitte Survey, 2022 | Chart: 2023 AI Index Report
success, 94% responded “important,” 5% said
“somewhat important,” and 1% answered “not 2%, Strongly Disagree / Disagree 16%, Neither Agree nor Disagree
1%, Unsure
important” (Figure 4.3.13).

Similarly, when asked whether they believe that AI

enhances performance and job satisfaction, 82%
responded “strongly agree/agree,” 16% said they
“neither agree nor disagree,” and only 2% selected
“strongly disagree/disagree” (Figure 4.3.14).

82%, Strongly Agree / Agree

Figure 4.3.14

Chapter 4 Preview 43
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

AI Investments and Implementation

a 9 percentage point decrease since 2021 and a 12
Outcomes
percentage point decrease since 2018, a significantly
In 2022, 76% of surveyed leaders reported
large portion of business leaders continue to express
expecting to increase AI investments in the next
interest in AI investment.
fiscal year (Figure 4.3.15). Although this represents

Expected AI Investment Increase in the Next Fiscal Year

Source: Deloitte Survey, 2022 | Chart: 2023 AI Index Report

80%
76%

60%
% of Respondents

40%

20%

0%
2018 2019 2020 2021 2022

Figure 4.3.15

Chapter 4 Preview 44
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

Figure 4.3.16 highlights the main outcomes that business leaders achieved by embracing AI solutions.7
The top outcome was lowered costs (37%), followed by improved collaboration across business
functions/organizations (34%) and having discovered valuable insights (34%).

Main Outcomes of AI Implementation, 2022

Source: Deloitte Survey, 2022 | Chart: 2023 AI Index Report
Lower Costs 37%
Improve Collaboration Across
Business Functions/Organizations 34%

Discover Valuable Insights 34%

Customize or Improve Product/Programs,
33%
Services, or Offers
Enter New Markets/Expand Services to 33%
New Constituents
Make Organizational Processes More E�cient 33%

Improve Decision-Making 32%

Create New Products/
Programs and Services 32%

Predict Demand 32%

Enable New Business/
Service Models 32%

Increase Revenue 31%

Activate the Potential of Existing Headcount
and/or Improve Talent Management 30%

Improve Constituent Engagement 30%

Anticipate Constituent Needs 28%

0% 10% 20% 30%

% of Respondents

Figure 4.3.16

7 Figure 4.3.16 is drawn from the chart in the Deloitte survey: “Outcomes—‘Achieved to a high degree.’”

Chapter 4 Preview 45
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

Challenges in Starting and Scaling AI Projects were proving business value (37%), lack of executive
The top three challenges that business leaders commitment (34%), and choosing the right AI
identified in terms of starting AI-related projects technologies (33%) (Figure 4.3.17).

Top Three Challenges in Starting AI Projects, 2022

Source: Deloitte Survey, 2022 | Chart: 2023 AI Index Report

Proving Business
37%
Value

Lack of Executive
34%
Commitment

Choosing the Right

33%
AI Technologies

0% 5% 10% 15% 20% 25% 30% 35%

% of Respondents

Figure 4.3.17

Chapter 4 Preview 46
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

The main barrier leaders faced in scaling existing AI initiatives was managing AI-related risks (50%), obtaining
more data or inputs to train a model (44%), and implementing AI technologies (42%) (Figure 4.3.18).

Main Barriers in Scaling AI Initiatives, 2022

Source: Deloitte Survey, 2022 | Chart: 2023 AI Index Report

Managing
50%
AI-Related Risks

Obtaining Needed Data

44%
or Input to Train Model

Implementing
42%
AI Technologies

Proving Business Value 40%

0% 10% 20% 30% 40% 50%

% of Respondents

Figure 4.3.18

Chapter 4 Preview 47
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

Earnings Calls Aggregate Trends

In the 2022 fiscal year, there were 268 earnings calls
The following subsection presents data from from Fortune 500 companies that mentioned AI-related
NetBase Quid, which uses natural language keywords (Figure 4.3.19). The number of such mentions
processing tools to analyze trends in corporate dropped from the previous year, when there were 306,
earnings calls. NetBase Quid analyzed all 2022 but has increased since 2018 when there were 225.
earnings calls from Fortune 500 companies,
identifying all mentions of “Artificial Intelligence,”
“AI,” “Machine Learning,” “ML,” and “deep learning.”

Number of Fortune 500 Earnings Calls Mentioning AI, 2018–22

Source: NetBase Quid, 2022 | Chart: 2023 AI Index Report

300

268
Number of Earnings Calls

200

100

0
2018 2019 2020 2021 2022

Figure 4.3.19

Chapter 4 Preview 48
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

Specific Themes advertising and marketing (8.8%) (Figure 4.3.20).

Mentions of AI in Fortune 500 earnings calls were Compared to 2018, some of the less prevalent
associated with a wide range of themes. In 2022, the AI-related themes in 2022 included deep learning
most cited themes were business integration (10.0%); (4.8%), autonomous vehicles (3.1%), and data
pricing and inventory management (8.8%); and storage and management (3.0%).

Themes for AI Mentions in Fortune 500 Earnings Calls, 2018 Vs. 2022
Source: NetBase Quid, 2022 | Chart: 2023 AI Index Report

9.96% (-15%)
Business Integration
11.74%
8.82% (+48%)
Pricing and Inventory Management
5.94%
8.82% (+204%)
Advertising and Marketing
2.90%
8.39% (+23%)
Process Automation
6.81%
7.40% (-7%)
Support Decision-Making
7.97%
7.11% (+69%)
Healthcare and Medical Practices
4.20%
6.26% (+73%)
Cloud Platforms
3.62%
5.26% (-21%)
Personalizing Customer Experience
6.67%
4.84% (-41%)
Deep Learning
8.26%
4.13% (+24%)
Edge Intelligence
3.33%
3.84% (+121%)
Nvidia AI Use Cases
1.74%
3.27% (+33%)
Revenue Growth
2.46%
3.13% (-47%)
Autonomous Vehicles
5.94%
2.99% (+37%)
Data Processing
2.17%
2.99% (-55%)
Data Storage and Management
6.67%
2.70% (+10%)
Adobe Experience
2.46%
2.42% (+734%)
Customer Support
0.29%
2.13% (-59%)
Azure Cognitive Services
5.22%
1.85% (-20%)
Data Center GPU
2.32%
1.28% (+47%)
Investments
0.87% 2022
1.00% (-62%) 2018
Nvidia RTX
2.61%
0.71% (-87%)
Digital Transformation
5.36%

0% 1% 2% 3% 4% 5% 6% 7% 8% 9% 10% 11% 12% 13% 14%

Theme Mentioned (% of Total)
Figure 4.3.20

Chapter 4 Preview 49
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

Narrative Highlight:
What Are Business Leaders Actually Saying About AI?

To better understand business attitudes that surround AI, it is worth

“We spent a ton of money
looking at AI-related excerpts from the Fortune 500 earnings calls.
on Cloud. We spend a ton of
For example, on the topic of business integration, companies often money on adding capabilities.
And over time, as you do it on
cite AI and machine learning (ML) use cases to reassure business
one platform, it all becomes
audiences of safer business practices, growing opportunities, more efficient. So, I think it’s
streamlining processes, and capability expansion. a lot of little things, but it adds
up with our base of people
and fixed cost, it adds up
“We spent $100 million building “Especially in the last year or so, the significantly over time. We’ve
certain risk and fraud systems field of robotics itself has actually been able to maintain our
so that when we process changed because with AI and headcount at a level we feel
payments on the consumer side, ML coming to the picture, there’s good about, and we think
losses are down $100 million to significant developments in the we can grow massively on
$200 million. Volume is way up. robotics field. So we think it’s a top of that without having to
That’s a huge benefit.” huge opportunity for us.” add lots of bodies to be able
– Jamie Dimon, CEO, JP Morgan – Raj Subramaniam, CEO, FedEx to do it.” – Peter Kern, CEO,
Chase & Co. (Q2 2022) (Q3 2022) Expedia Group (Q4 2022)

In terms of process automation, business leaders emphasize the ability of AI tools to accelerate
productivity gains and to deliver a better customer experience.

“We continue to drive “We have improved the “In September, we opened a next-
the use of automation experience for customers by gen fulfillment center in Illinois.
and artificial applying artificial intelligence This 1.1 million square foot facility
intelligence to drive to match them with an expert features robotics, machine learning,
productivity gains to who is right for their specific and automated storage, resulting in
help offset inflationary situation and to deliver insights increased productivity and a better
pressures.” – Jim Davis, to experts so they can provide service for our customers at faster
CEO, Quest Diagnostics excellent service.” – Sasan delivery times.” – John David, CFO,
(Q4 2022) Goodarzi, CEO, Intuit (Q2 2022) Walmart (Q3 2022)

Chapter 4 Preview 50
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

Narrative Highlight:
What Are Business Leaders Actually Saying About AI?
(cont’d)

The conversation surrounding pricing and inventory management saw companies reassuring business
audiences on how their use of AI would improve their operational strength, especially in environments of
high inflation and supply chain challenges.

“We are … continuing to refine and invest “Our teams are utilizing technology, innovative data analytics
in machine learning tools that will allow for and AI to forecast supply chain lead times and changes
more sophisticated competitive pricing in market demand to ensure optimal levels. These actions
and greater automation at scale.” along with our pricing initiatives positively impacted our
– Adrian Mitchell, CFO, Macy’s (Q3 2022) gross margin in the second quarter.”
– Bert Nappier, CFO, Genuine Parts Company (Q3 2022)

There is also a vibrant discussion about the ways in which AI can change healthcare and medical
practices, more specifically to reduce costs, improve the patient experience, and better serve clinicians.

“[Using] machine “I’d like to highlight productivity efforts in our preauthorization process where
learning and robotics, we’re leveraging an in-house artificial intelligence solution to automatically
we can now resolve match incoming faxes to the correct authorization requests. This solution
a wide range of creates administrative efficiencies across millions of inbound images. We are
prescription drug also scaling this solution to multiple business units such as pharmacy and
claims which previously are also expanding the application of this type of AI to provide decision
required the attention of support to clinicians, which will result in improvements to authorization
our pharmacists, freeing turnaround times, reduction in friction for providers and creating a better
them up to spend time member experience.” – Bruce Broussard, CEO, Humana (Q3 2022)
with patients. This
advanced approach
reduces overall cost
and improves the “We continue to see opportunities across [the software and analytics] segment
patient experience.” as payers, providers, and partners take advantage of our high ROI solutions and
– Karen Lynch, CEO, realize the benefits of our data, AI models, and workflow capabilities.”
CVS Health (Q2 2022) – Neil de Crescenzo, CEO, UnitedHealth Group (Q2 2022)

Chapter 4 Preview 51
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.3 Corporate Activity

Sentiment Analysis sentiment associated with mentions of AI has been

NetBase Quid also runs the AI-related text of Fortune overwhelmingly positive (Figure 4.3.21). Mentions
500 earnings calls through a sentiment analysis of AI were rarely negative, suggesting that large
machine-learning algorithm that identifies whether businesses tend to have positive associations when it
the sentiment associated with the mention of AI is comes to AI tools.
positive, mixed, or negative . Overall, since 2018, the 8

Sentiment Summary Distribution for AI Mentions in Fortune 500 Earnings Calls by Publication Date, 2018–22
Source: NetBase Quid, 2022 | Chart: 2023 AI Index Report

100% 1%% 1%% 0%%

2%%
1%%
2%%

3%% 3%%
1%% 1%% 1%%
0%%
1%%
2%%
1%% 1%% 1%%
1%% 2%%
3%%

13%%
18%%
19%%

16%%
15%%

16%%
14%%
13%%

15%%

20%%

18%%
22%%
23%%
18%%
17%%

19%%
12%%
21%%

17%%
17%%

87%%

86%%
86%%

85%%
84%%

84%%

84%%
84%%
80%
81%%
81%%

81%%
81%%

81%%

80%%

80%%
79%%

79%%
77%%
77%%

76%%
Sentiment Summary

60%

40%

Negative
20% Mixed
Positive

0%
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4
2018 2019 2020 2021 2022

Figure 4.3.21

8 Chapter 2 of the 2023 AI Index highlights trends in the performance of sentiment analysis algorithms.

Chapter 4 Preview 52
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.4 Robot Installations

Given that robots are frequently deployed with AI-based software technologies, it is possible to gain insights on AI-ready infrastructure
being deployed in the real world by tracking the installation of industrial robots. Data in this section comes from the International
Federation of Robotics (IFR), an international nonprofit organization that works to promote, strengthen, and protect the robotics
industry. Every year the IFR releases the World Robotics Report, which tracks global trends in installations of robots.9

4.4 Robot Installations

Aggregate Trends
2021 saw a rebound in the total number of worldwide
The following subsection includes data on the robot installations. The 517,000 industrial robots
installation and operation of industrial robots, installed in 2021 represented a 31.3% increase from
which are defined as an “automatically controlled, 2020 and a 211.5% increase since 2011 (Figure 4.4.1).
reprogrammable, multipurpose manipulator,
programmable in three or more axes, which can be
either fixed in place or mobile for use in industrial
automation applications.”

Number of Industrial Robots Installed in the World, 2011–21

Source: International Federation of Robotics (IFR), 2022 | Chart: 2023 AI Index Report

517
500
Number of Industrial Robots Installed (in Thousands)

400

300

200

100

0
2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

Figure 4.4.1

9 Due to the timing of the IFR’s survey, the most recent data is from 2021.

Chapter 4 Preview 53
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.4 Robot Installations

The worldwide operational stock of industrial robots from 3,035,000 in 2020. In the last decade, the
also continues to steadily increase year over year number of industrial robots being installed and the
(Figure 4.4.2). The total number of operational number being used have both steadily increased.
industrial robots jumped 14.6% to 3,477,000 in 2021,

Operational Stock of Industrial Robots in the World, 2011–21

Source: International Federation of Robotics (IFR), 2022 | Chart: 2023 AI Index Report

3,477
3,500

3,000
Number of Industrial Robots (in Thousands)

2,500

2,000

1,500

1,000

500

0
2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

Figure 4.4.2

Chapter 4 Preview 54
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.4 Robot Installations

Industrial Robots: Traditional Vs. scalable than traditional robots, and are capable of
Collaborative Robots iterative learning.
A distinction can be drawn between traditional
robots that work for humans and collaborative In 2017, only 2.8% of all newly installed industrial
robots that are designed to work with humans. robots were collaborative (Figure 4.4.3). As of 2021,
Recently, the robotics community has been excited that number increased to 7.5%. Although traditional
about the potential of collaborative robots given industrial robots still lead new installations, the
that they can be safer, more flexible, and more number of collaborative robots is slowly increasing.

Number of Industrial Robots Installed in the World by Type, 2017–21

Source: International Federation of Robotics (IFR), 2022 | Chart: 2023 AI Index Report

Traditional 517
500 Collaborative 39
Number of Industrial Robots Installed (in Thousands)

424
400 394
400 391

300

478

200 389 405

370 368

100

0
2017 2018 2019 2020 2021

Figure 4.4.3

Chapter 4 Preview 55
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.4 Robot Installations

By Geographic Area the amount installed by Japan (47,200) and 7.7

Country-level data on robot installations can illustrate times the amount installed by the United States
which countries are prioritizing the integration of (35,000) (Figure 4.4.4). The countries with the next
robots into their economy. In 2021, China installed most installations were South Korea (31,100) and
the most industrial robots, with 268,200, 5.7 times Germany (23,800).

Number of Industrial Robots Installed by Country, 2021

Source: International Federation of Robotics (IFR), 2022 | Chart: 2023 AI Index Report

China 268.20

Japan 47.20

United States 35.00

South Korea 31.10

Germany 23.80

Italy 14.10

Taiwan 9.60

France 5.90

Mexico 5.40

India 4.90

Canada 4.30

Thailand 3.90

Singapore 3.50

Spain 3.40

Poland 3.30

0 30 60 90 120 150 180 210 240 270

Number of Industrial Robots Installed (in Thousands)

Figure 4.4.4

Chapter 4 Preview 56
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.4 Robot Installations

In 2013, China overtook Japan as the nation installing only widened. In 2013, Chinese industrial robot
the most industrial robots (Figure 4.4.5). Since then, installations represented 20.8% of the world’s share,
the gap between the total number of industrial robots whereas in 2021, they represented 51.8%.
installed by China and the next-nearest nation has

Number of New Industrial Robots Installed in Top Five Countries, 2011–21

Source: International Federation of Robotics (IFR), 2022 | Chart: 2023 AI Index Report

268, China

250
Number of Industrial Robots Installed (in Thousands)

200

150

100

50 47, Japan
35, United States
31, South Korea
24, Germany

0
2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

Figure 4.4.5

Chapter 4 Preview 57
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.4 Robot Installations

China consolidated its dominance in industrial robotics in 2021, the first year in which the country installed
more industrial robots than the rest of the world combined (Figure 4.4.6).

Number of Industrial Robots Installed (China Vs. Rest of the World), 2016–21
Source: International Federation of Robotics (IFR), 2022 | Chart: 2023 AI Index Report

268, China

250 249, Rest of the World

Number of Industrial Robots Installed (in Thousands)

200

150

100

50

0
2016 2017 2018 2019 2020 2021

Figure 4.4.6

Chapter 4 Preview 58
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.4 Robot Installations

Figure 4.4.7 shows the annual growth rate of of industrial robot installations. The countries that
industrial robot installations from 2020 to 2021 by reported the highest growth rates were Canada
country. Virtually every country surveyed by the (66%), Italy (65%), and Mexico (61%).
IFR reported a yearly increase in the total number

Annual Growth Rate of Industrial Robots Installed by Country, 2020 Vs. 2021
Source: International Federation of Robotics (IFR), 2022 | Chart: 2023 AI Index Report

Canada 66%

Italy 65%

Mexico 61%

Poland 56%

India 54%

China 51%

Thailand 36%

Taiwan 31%

Japan 22%

United States 14%

France 11%

Germany 6%

South Korea 2%

Spain 1%

Singapore -35%

−40% −30% −20% −10% 0% 10% 20% 30% 40% 50% 60% 70%
Annual Growth Rate of Industrial Robots Installed

Figure 4.4.7

Chapter 4 Preview 59
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.4 Robot Installations

Narrative Highlight:
Country-Level Data on Service Robotics
Another important class of robots are service robots, Service Robots in Medicine
Source: UL Solutions, 2022
which the ISO defines as a robot “that performs
useful tasks for humans or equipment excluding
industrial automation applications.”10 Figure 4.4.8
is an example of a robot being used in medicine,
Figure 4.4.9 illustrates how a robot can help with
professional cleaning, and Figure 4.4.10 shows a
robot designed for maintenance and inspection.

Figure 4.4.8

Service Robots in Professional Cleaning Service Robots in Maintenance and Inspection

Source: This Week in FM, 2021 Source: Robotnik, 2022

Figure 4.4.9 Figure 4.4.10

10 A more detailed definition can be accessed here.

Chapter 4 Preview 60
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.4 Robot Installations

Narrative Highlight:
Country-Level Data on Service Robotics (cont’d)
Compared to 2020, 2021 saw a higher number of professional service robots installed in the world
for several key application areas, including hospitality, medical robotics, professional cleaning, and
transportation and logistics (Figure 4.4.11). The category that registered the greatest year-over-year
increase was transportation and logistics: In 2021, 1.5 times the number of such service robots were
installed as in 2020.

Number of Professional Service Robots Installed in the World by Application Area, 2020 Vs. 2021
Source: International Federation of Robotics (IFR), 2022 | Chart: 2023 AI Index Report

8
Agriculture
8

20
Hospitality
11

15
Medical Robotics
12

13
Professional Cleaning 2021
10
2020

50
Transportation and Logistics
34

0 5 10 15 20 25 30 35 40 45 50
Number of Professional Service Robots Installed (in Thousands)
Figure 4.4.11

Chapter 4 Preview 61
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.4 Robot Installations

Narrative Highlight:
Country-Level Data on Service Robotics (cont’d)
As of 2022, the United States has the greatest number of professional service robot manufacturers,
roughly 2.16 times as many as the next nation, China. Other nations with significant numbers of robot
manufacturers include Germany (91), Japan (66), and France (54) (Figure 4.4.12).

Number of Professional Service Robot Manufacturers in Top Countries by Type of Company, 2022
Source: International Federation of Robotics (IFR), 2022 | Chart: 2023 AI Index Report

225

29
200
Number of Professional Service Robot Manufacturers

150

104
100 194 Startups
91
Incumbents
Unknown
66
54 52
50 94 47
39 39
79
61
49 44 44
34 35

0
3
2

United States China Germany Japan France Russia South Korea Switzerland Canada

Figure 4.4.12

Chapter 4 Preview 62
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.4 Robot Installations

Sectors and Application Types (137,000), followed by automotive (119,000) (Figure

4.4.13). Each of the highlighted sectors has recorded
On a global level, the sector that saw the greatest increases in the total number of industrial robot
amount of robot installations was electrical/electronics installations since 2019.

Number of Industrial Robots Installed in the World by Sector, 2019–21

Source: International Federation of Robotics (IFR), 2022 | Chart: 2023 AI Index Report

52
All Others 37
30

119
Automotive 84
102

137
Electrical/Electronics 110
89

15
Food 12
11

64
Metal and Machinery 44
52

24
Plastic and Chemical Products 19
18 2021
107 2020
Unspeci ed 87 2019
87

0 20 40 60 80 100 120 140

Number of Industrial Robots Installed (in Thousands)

Figure 4.4.13

Chapter 4 Preview 63
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.4 Robot Installations

Robots can also be deployed in a wide range of 230,000 industrial robots were installed for handling
applications, from assembling to dispensing and functions, 2.4 times more than for welding (96,000)
handling. Figure 4.4.14 illustrates how the application and 3.7 times more than for assembling (62,000).
of industrial robots has changed since 2021. Handling Every application category, with the exception
continues to be the application case toward which of dispensing and processing, saw more robot
the most industrial robots are deployed. In 2021, installations in 2021 than in 2019.

Number of Industrial Robots Installed in the World by Application, 2019–21

Source: International Federation of Robotics (IFR), 2022 | Chart: 2023 AI Index Report

62
Assembling 50
40

32
Cleanroom 32
26

11
Dispensing 8
12

230
Handling 169
177

7
Processing 5
7 2021
2020
80
2019
Unspeci ed 60
55

96
Welding 70
74

0 20 40 60 80 100 120 140 160 180 200 220 240

Number of Industrial Robots Installed (in Thousands)

Figure 4.4.14

Chapter 4 Preview 64
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.4 Robot Installations

China Vs. United States and metal and machinery (34,000) (Figure 4.4.15).
The Chinese industrial sectors that installed the Every industrial sector in China recorded a greater
greatest number of industrial robots in 2022 were number of robot installations in 2021 than in 2019.
electrical/electronics (88,000), automotive (62,000),

Number of Industrial Robots Installed in China by Sector, 2019–21

Source: International Federation of Robotics (IFR), 2022 | Chart: 2023 AI Index Report

29
All Others 21
12

62
Automotive 31
32

88
Electrical/Electronics 64
42

4
Food 3
3

34
Metal and Machinery 22
22

1
Pharma/Cosmetics 1
1

6
Rubber and Plastics 5
4 2021
43
2020
Unspeci ed 30 2019
31

0 10 20 30 40 50 60 70 80 90
Number of Industrial Robots Installed (in Thousands)

Figure 4.4.15

Chapter 4 Preview 65
 Artificial Intelligence Chapter 4: The Economy
Index Report 2023 4.4 Robot Installations

The automotive industry installed the greatest number of industrial robots in the United States in 2021,
although installation rates for that sector decreased year over year (Figure 4.4.16). However, other sectors like
food, along with plastic and chemical products, saw year-over-year increases in robot installations.

Number of Industrial Robots Installed in the United States by Sector, 2019–21

Source: International Federation of Robotics (IFR), 2022 | Chart: 2023 AI Index Report

4.50
All Others 2.60
3.50

9.80
Automotive 10.50
13.00

2.90
Electrical/Electronics 3.70
3.50

3.40
Food 2.70
2.20

3.80
Metal and Machinery 2.30
3.80

3.50
Plastic and Chemical Products 2.60
2.50 2021
7.10 2020
Unspeci ed 6.30 2019
5.00

0 2 3 5 6 8 9 11 12 14
Number of Industrial Robots Installed (in Thousands)

Figure 4.4.16

Chapter 4 Preview 66
 Artificial Intelligence Appendix
Index Report 2023 Chapter 4: The Economy

Appendix
Lightcast job postings is the Job Openings and Labor Turnover
Survey (JOLTS) program, conducted by the Bureau
Prepared by Scott Bingham, Julia Nania, Layla O’Kane, of Labor Statistics. Based on comparisons between
and Bledi Taska JOLTS and Lightcast, the labor market demand
Lightcast delivers job market analytics that empower captured by Lightcast data represents over 99% of
employers, workers, and educators to make data- the total labor demand. Jobs not posted online are
driven decisions. The company’s artificial intelligence usually in small businesses (the classic example being
technology analyzes hundreds of millions of job postings the “Help Wanted” sign in a restaurant window) and
and real-life career transitions to provide insight union hiring halls.
into labor market patterns. This real-time strategic Measuring Demand for AI
intelligence offers crucial insights, such as what jobs are In order to measure the demand by employers of AI
most in demand, the specific skills employers need, and skills, Lightcast uses its skills taxonomy of over 31,000
the career directions that offer the highest potential for skills. The list of AI skills from Lightcast data are shown
workers. For more information, visit www.lightcast.io. below, with associated skill clusters. While some skills
Job Posting Data are considered to be in the AI cluster specifically,
To support these analyses, Lightcast mined its dataset for the purposes of this report, all skills below were
of millions of job postings collected since 2010. considered AI skills. A job posting was considered an
Lightcast collects postings from over 51,000 online job AI job if it mentioned any of these skills in the job text.
sites to develop a comprehensive, real-time portrait Artificial Intelligence: AIOps (Artificial Intelligence for
of labor market demand. It aggregates job postings, IT Operations), Applications of Artificial Intelligence,
removes duplicates, and extracts data from job postings Artificial General Intelligence, Artificial Intelligence,
text. This includes information on job title, employer, Artificial Intelligence Development, Artificial
industry, and region, as well as required experience, Intelligence Markup Language (AIML), Artificial
education, and skills. Intelligence Systems, Azure Cognitive Services,
Job postings are useful for understanding trends in Baidu, Cognitive Automation, Cognitive Computing,
the labor market because they allow for a detailed, Computational Intelligence, Cortana, Expert Systems,
real-time look at the skills employers seek. To assess Intelligent Control, Intelligent Systems, Interactive
the representativeness of job postings data, Lightcast Kiosk, IPSoft Amelia, Knowledge-Based Configuration,
conducts a number of analyses to compare the Knowledge-Based Systems, Multi-Agent Systems,
distribution of job postings to the distribution of official Open Neural Network Exchange (ONNX), OpenAI
government and other third-party sources in the United Gym, Reasoning Systems, Soft Computing, Syman,
States. The primary source of government data on U.S. Watson Conversation, Watson Studio, Weka

Chapter 4 Preview 67
 Artificial Intelligence Appendix
Index Report 2023 Chapter 4: The Economy

Autonomous Driving: Advanced Driver Assistance Machine Learning: AdaBoost, Apache MADlib,
Systems, Autonomous Cruise Control Systems, Apache Mahout, Apache SINGA, Apache Spark,
Autonomous System, Autonomous Vehicles, Guidance Association Rule Learning, Automated Machine
Navigation and Control Systems, Light Detection and Learning, Autonomic Computing, AWS SageMaker,
Ranging (LiDAR), OpenCV, Path Analysis, Path Finding, Azure Machine Learning, Boosting, CHi-Squared
Remote Sensing, Unmanned Aerial Systems (UAS) Automatic Interaction Detection (CHAID),
Classification And Regression Tree (CART), Cluster
Natural Language Processing (NLP): Amazon Textract,
Analysis, Collaborative Filtering, Confusion Matrix,
ANTLR, BERT (NLP Model), Chatbot, Computational
Cyber-Physical Systems, Dask (Software), Data
Linguistics, DeepSpeech, Dialog Systems, fastText,
Classification, DBSCAN, Decision Models, Decision
Fuzzy Logic, Handwriting Recognition, Hugging
Tree Learning, Dimensionality Reduction, Dlib
Face (NLP Framework), HuggingFace Transformers,
(C++ Library), Ensemble Methods, Evolutionary
Intelligent Agent, Intelligent Software Assistant,
Programming, Expectation Maximization Algorithm,
Intelligent Virtual Assistant, Kaldi, Latent Dirichlet
Feature Engineering, Feature Extraction, Feature
Allocation, Lexalytics, Machine Translation, Microsoft
Learning, Feature Selection, Gaussian Process,
LUIS, Natural Language Generation, Natural Language
Genetic Algorithm, Google AutoML, Google Cloud
Processing, Natural Language Processing Systems,
ML Engine, Gradient Boosting, H2O.ai, Hidden
Natural Language Programming, Natural Language
Markov Model, Hyperparameter Optimization,
Toolkits, Natural Language Understanding, Natural
Inference Engine, K-Means Clustering, Kernel
Language User Interface, Nearest Neighbour
Methods, Kubeflow, LIBSVM, Machine Learning,
Algorithm, OpenNLP, Optical Character Recognition
Machine Learning Algorithms, Markov Chain, Matrix
(OCR), Screen Reader, Semantic Analysis, Semantic
Factorization, Meta Learning, Microsoft Cognitive
Interpretation for Speech Recognition, Semantic
Toolkit (CNTK), MLflow, MLOps (Machine Learning
Parsing, Semantic Search, Sentiment Analysis,
Operations), mlpack (C++ Library), Naive Bayes,
Seq2Seq, Speech Recognition, Speech Recognition
Perceptron, Predictionio, PyTorch (Machine Learning
Software, Statistical Language Acquisition, Text Mining,
Library), Random Forest Algorithm, Recommendation
Tokenization, Voice Interaction, Voice User Interface,
Engine, Recommender Systems, Reinforcement
Word Embedding, Word2Vec Models
Learning, Scikit-learn (Machine Learning Library),
Neural Networks: Apache MXNet, Artificial Neural Semi-Supervised Learning, Soft Computing, Sorting
Networks, Autoencoders, Caffe, Caffe2, Chainer, Algorithm, Supervised Learning, Support Vector
Convolutional Neural Networks, Cudnn, Deep Learning, Machine, Test Datasets, Torch (Machine Learning),
Deeplearning4j, Keras (Neural Network Library), Long Training Datasets, Transfer Learning, Unsupervised
Short-Term Memory (LSTM), OpenVINO, PaddlePaddle, Learning, Vowpal Wabbit, Xgboost
Pybrain, Recurrent Neural Network (RNN), TensorFlow

Chapter 4 Preview 68
 Artificial Intelligence Appendix
Index Report 2023 Chapter 4: The Economy

Robotics: Advanced Robotics, Cognitive Robotics, intelligence, computer vision, image processing,
Motion Planning, Nvidia Jetson, Robot Framework, deep learning, TensorFlow, Pandas (software), and
Robot Operating Systems, Robotic Automation OpenCV, among others.
Software, Robotic Liquid Handling Systems, Robotic
Skill groupings are derived by expert taxonomists
Programming, Robotic Systems, Servomotor, SLAM
through a similarity-index methodology that
Algorithms (Simultaneous Localization and Mapping)
measures skill composition at the industry level.
Visual Image Recognition: 3D Reconstruction, Activity LinkedIn’s industry taxonomy and their corresponding
Recognition, Computer Vision, Contextual Image NAICS codes can be found here.
Classification, Digital Image Processing, Eye Tracking,
Skills Genome
Face Detection, Facial Recognition, Image Analysis,
For any entity (occupation or job, country, sector,
Image Matching, Image Processing, Image Recognition,
etc.), the skill genome is an ordered list (a vector) of
Image Segmentation, Image Sensor, Imagenet,
the 50 “most characteristic skills” of that entity. These
Machine Vision, Motion Analysis, Object Recognition,
most characteristic skills are identified using a TF-IDF
OmniPage, Pose Estimation, RealSense
algorithm to identify the most representative skills of
the target entity, while down-ranking ubiquitous skills
LinkedIn that add little information about that specific entity
Prepared by Murat Erer and Akash Kaura (e.g., Microsoft Word).

Country Sample TF-IDF is a statistical measure that evaluates how

Included countries represent a select sample of eligible representative a word (in this case a skill) is to a
countries with at least 40% labor force coverage by selected entity). This is done by multiplying two
LinkedIn and at least 10 AI hires in any given month. metrics:
China and India were included in this sample because
1. The term frequency of a skill in an entity (TF).
of their increasing importance in the global economy,
but LinkedIn coverage in these countries does not 2. The logarithmic inverse entity frequency of the
reach 40% of the workforce. Insights for these countries skill across a set of entities (IDF). This indicates
may not provide as full a picture as other countries, and how common or rare a word is in the entire entity
should be interpreted accordingly. set. The closer IDF is to 0, the more common the
word.
Skills (and AI Skills)
LinkedIn members self-report their skills on their So if the skill is very common across LinkedIn entities,
LinkedIn profiles. Currently, more than 38,000 distinct, and appears in many job or member descriptions, the
standardized skills are identified by LinkedIn. These IDF will approach 0. If, on the other hand, the skill
have been coded and classified by taxonomists at is unique to specific entities, the IDF will approach
LinkedIn into 249 skill groupings, which are the skill 1. More details are available at LinkedIn’s Skills
groups represented in the dataset. The top skills that Genome and LinkedIn-World Bank Methodology.
make up the AI skill grouping are machine learning,
natural language processing, data structures, artificial

Chapter 4 Preview 69
 Artificial Intelligence Appendix
Index Report 2023 Chapter 4: The Economy

AI Skills Penetration perform the job. Skills penetration is used as a signal

The aim of this indicator is to measure the intensity of AI for whether AI skills are prevalent in an occupation
skills in an entity (a particular country, industry, gender, representative in any sector where the occupation
etc.) through the following methodology: representative may exist. Examples of such
occupations include (but are not limited to): machine
• Compute frequencies for all self-added skills by
learning engineer, artificial intelligence specialist,
LinkedIn members in a given entity (occupation,
data scientist, computer vision engineer, etc.
industry, etc.) in 2015–2021.
AI Talent
• Re-weight skill frequencies using a TF-IDF model
A LinkedIn member is considered AI talent if they
to get the top 50 most representative skills in
have explicitly added AI skills to their profile and/or
that entity. These 50 skills compose the “skill
they are occupied in an AI occupation representative.
genome” of that entity.
The counts of AI talent are used to calculate talent
• Compute the share of skills that belong to the concentration metrics. For example, to calculate
AI skill group out of the top skills in the selected the country level AI talent concentration, we use
entity. the counts of AI talent at the country level vis-a-vis
the counts of LinkedIn members in the respective
Interpretation: The AI skill penetration rate signals
countries.
the prevalence of AI skills across occupations, or the
intensity with which LinkedIn members utilize AI skills Relative AI Skills Penetration
in their jobs. For example, the top 50 skills for the To allow for skills penetration comparisons across
occupation of engineer are calculated based on the countries, the skills genomes are calculated and a
weighted frequency with which they appear in LinkedIn relevant benchmark is selected (e.g., global average).
members’ profiles. If four of the skills that engineers A ratio is then constructed between a country’s and
possess belong to the AI skill group, this measure the benchmark’s AI skills penetrations, controlling for
indicates that the penetration of AI skills is estimated to occupations.
be 8% among engineers (i.e., 4/50). Interpretation: A country’s relative AI skills
Jobs or Occupations penetration of 1.5 indicates that AI skills are 1.5 times
LinkedIn member titles are standardized and grouped as frequent as in the benchmark, for an overlapping
into approximately 15,000 occupations. These are set of occupations.
not sector- or country-specific. These occupations Global Comparison
are further standardized into approximately For cross-country comparison, we present the
3,600 occupation representatives. Occupation relative penetration rate of AI skills, measured as
representatives group occupations with a common role the sum of the penetration of each AI skill across
and specialty, regardless of seniority. occupations in a given country, divided by the
AI Jobs and Occupations average global penetration of AI skills across the
An “AI” job (technically, occupation representative) overlapping occupations in a sample of countries.
is an occupation representative that requires AI skills to

Chapter 4 Preview 70
 Artificial Intelligence Appendix
Index Report 2023 Chapter 4: The Economy

Interpretation: A relative penetration rate of 2 means Interpretation: A country’s “Relative AI Skills

that the average penetration of AI skills in that country Penetration Across Genders” for women of 1.5 means
is two times the global average across the same set of that female members in that country are 1.5 times
occupations. more likely to list AI skills than the average member in
all countries pooled together across the same set of
Global Comparison: By Industry
occupations that exist in the country.
The relative AI skills penetration by country for industry
provides an in-depth sectoral decomposition of AI skill Relative AI Hiring Index
penetration across industries and sample countries. 
LinkedIn Hiring Rate or Overall Hiring Rate
is a measure of hires normalized by LinkedIn
Interpretation: A country’s relative AI skill penetration
membership. It is computed as the percentage of
rate of 2 in the education sector means that the average
LinkedIn members who added a new employer
penetration of AI skills in that country is two times the
in the same period the job began, divided by
global average across the same set of occupations in
the total number of LinkedIn members in the
that sector.
corresponding location.
Global Comparison: By Gender
AI Hiring Rate is computed following the overall
The “Relative AI Skills Penetration by Gender” metric
hiring rate methodology, but only considering
provides a cross-country comparison of AI skill
members classified as AI talent.
penetrations within each gender, comparing countries’
male or female AI skill penetrations to the global 
Relative AI Hiring Index is the pace of change in
average of the same gender. Since the global averages AI Hiring Rate normalized by the pace of change
are distinct for each gender, this metric should only be in Overall Hiring Rate, providing a picture of
used to compare country rankings within each gender, whether hiring of AI talent is growing at a higher,
and not for cross-gender comparisons within countries. equal, or lower rate than overall hiring in a
market. The relative AI Hiring Index is equal to
Interpretation: A country’s AI skills penetration for
1.0 when AI hiring and overall hiring are growing
women of 1.5 means that female members in that
at the same rate year on year.
country are 1.5 times more likely to list AI skills than the
average female member in all countries pooled together Interpretation: Relative AI Hiring Index shows how
across the same set of occupations that exist in the fast each country is experiencing growth in AI talent
country/gender combination. hiring relative to growth in overall hiring in the country.
Global Comparison: Across Gender A ratio of 1.2 means the growth in AI talent hiring has
The “Relative AI Skills Penetration Across Genders” outpaced the growth in overall hiring by 20%.
metric allows for cross-gender comparisons within
and across countries globally, since we compare the
countries’ male and female AI skill penetrations to the
same global average regardless of gender.

Chapter 4 Preview 71
 Artificial Intelligence Appendix
Index Report 2023 Chapter 4: The Economy

Changelog From Methodology Included in Last Year’s AI Index

1. LinkedIn ramped a new version of its industry taxonomy (see details here).

a. This has resulted in changes to our top level five key industries. We have made the full-time series
available for each industry (as with prior years).

i. “Software & IT Services” industry evolved into a wider “Technology, Information and Media,” which
encompasses media and telecommunications as well as other sub-industries.

ii. Former “Hardware & Networking” industry does not exist in the new taxonomy, so we introduced
“Professional Services” industry as the fifth industry in scope which contains a high concentration of AI
talent.

iii. Remaining “Education,” “Manufacturing,” and “Financial Services” (formerly known as “Finance”)
also had updates in their coverage resulting from the inclusion of more granular sub-industries.

b. This also resulted in minor changes in magnitudes for some metrics since the distinct number of
industries, as well as the distinct number of AI occupations defined within each country-industry pair have
changed:

i. We define AI occupations (occupation representatives that require AI skills to perform the job) and
the respective definition of AI Talent at Country-Industry level. For example, data engineers working
in the technology, information, and media industry in Germany may be identified as holding an AI
occupation, whereas data engineers working in the construction industry in the United Arab Emirates
may not be identified as AI Talent. Following the introduction of a more granular industry taxonomy
with improved accuracy, our AI Talent identifications have been improved, and results have been
reflected to the entirety of time series for each relevant metric.

ii. The following metrics have been impacted by this change in industry taxonomy: AI Talent
Concentrations, and Relative AI Hiring Rates. No directional changes were observed, only minor
changes in magnitudes.

2. We introduced a methodology change into Relative Skills Penetration metrics:

a. In the past, the data used to calculate these metrics were limited to top five industries with the highest
AI skill penetration globally: “Software & IT Services,” “Hardware & Networking,” “Manufacturing,”
“Education,” and “Finance” industries. This year we updated our coverage to all industries.

Chapter 4 Preview 72
 Artificial Intelligence Appendix
Index Report 2023 Chapter 4: The Economy

NetBase Quid subsidiary, out of business) throughout the world.

The investment data includes private investments,
Prepared by Bill Valle and Nicole Seredenko M&A, public offerings, minority stakes made by PE/
NetBase Quid delivers AI-powered consumer and VCs, corporate venture arms, governments, and
market intelligence to enable business reinvention institutions both within and outside the United States.
in a noisy and unpredictable world. The software Some data is simply unreachable—for instance, when
applies artificial intelligence to reveal patterns in large, investors’ names or funding amounts are undisclosed.
unstructured datasets and to generate visualizations NetBase Quid embeds Capital IQ data as a default
that enable users to make smart, data-driven decisions and adds in data from Crunchbase for the data
accurately, quickly, and efficiently. NetBase Quid uses points that are not captured in Capital IQ. This not
Boolean query to search for focus areas, topics, and only yields comprehensive and accurate data on
keywords within social media, news, forums and blogs, all global organizations, but it also captures early-
companies, and patents data sources, as well as other stage startups and funding events data. Company
custom datasets. NetBase Quid then visualizes these information is updated on a weekly basis.
data points based on the semantic similarity.
Earnings Calls
Search, Data Sources, and Scope NetBase Quid leverages earnings call transcript
Over 8 million global public and private company profiles data embedded from Seeking Alpha. For this report,
from multiple data sources are indexed in order to NetBase Quid has analyzed mentions of AI-related
search across company descriptions, while filtering and keywords across all earnings call transcripts from
including metadata ranging from investment information Fortune 500 companies from January 2018 through
to firmographic information, such as founded year, HQ December 2022. New earnings call transcript data is
location, and more. Company information is updated updated in NetBase Quid on the 1st and 15th of every
on a weekly basis. The NetBase Quid algorithm reads month.
a big amount of text data from each document to
Search Parameters
make links between different documents based on
Boolean query is used to search for focus areas,
their similar language. This process is repeated at an
topics, and keywords within the archived company
immense scale, which produces a network with different
database, within their business descriptions and
clusters identifying distinct topics or focus areas. Trends
websites. We can filter out the search results by
are identified based on keywords, phrases, people,
HQ regions, investment amount, operating status,
companies, and institutions that NetBase Quid identifies,
organization type (private/public), and founding
and the other metadata that is put into the software.
year. NetBase Quid then visualizes these companies
Data by semantic similarity. If there are more than 7,000
Companies companies from the search result, NetBase Quid
Organization data is embedded from Capital IQ and selects the 7,000 most relevant companies for
Crunchbase. These companies include all types of visualization based on the language algorithm.
companies (private, public, operating, operating as a

Chapter 4 Preview 73
 Artificial Intelligence Appendix
Index Report 2023 Chapter 4: The Economy

Boolean Search: “artificial intelligence” or “AI” or

McKinsey & Company
“machine learning” or “deep learning”
Data used in the Corporate Activity-Industry
Companies:
Adoption section was sourced from the McKinsey
• Global AI and ML companies that have received Global Survey “The State of AI in 2022—and a Half
investments (private, IPO, M&A) from January 1, Decade in Review.”
2013, to December 31, 2022.
The online survey was in the field from May 3, 2022,
• Global AI and ML companies that have received to May 27, 2022, and from August 15, 2022, to
over $1.5M for the last 10 years (January 1, 2013, August 17, 2022, and garnered responses from 1,492
to December 31, 2022): 7,000 out of 7,500 participants representing a full range of regions,
companies have been selected through NetBase industries, company sizes, functional specialties,
Quid’s relevance algorithm. and tenures. Of those respondents, 744 said their
organization had adopted AI in at least one function
Target Event Definitions
and were asked questions about their organization’s
• Private investments: A private placement is a AI use. To adjust for differences in response rates,
private sale of newly issued securities (equity or the data is weighted by the contribution of each
debt) by a company to a selected investor or a respondent’s nation to global GDP.
selected group of investors. The stakes that buyers
The AI Index also considered data from previous
take in private placements are often minority
iterations of the survey. More specifically, the AI
stakes (under 50%), although it is possible to take
index made use of data from:
control of a company through a private placement
as well, in which case the private placement would The State of AI in 2021

be a majority stake investment. The State of AI in 2020

Global AI Survey: AI Proves Its Worth,
• Minority investment: These refer to minority
But Few Scale Impact (2019)
stake acquisitions in NetBase Quid, which take
AI Adoption Advances, But Foundational Barriers
place when the buyer acquires less than 50% of
Remain (2018)
the existing ownership stake in entities, asset
products, and business divisions.

• M&A: This refers to a buyer acquiring more than

50% of the existing ownership stake in entities,
asset products, and business divisions.

Chapter 4 Preview 74
 Artificial Intelligence Appendix
Index Report 2023 Chapter 4: The Economy

GitHub To obtain a global view of how AI is transforming

organizations, Deloitte surveyed 2,620 global
Data on the effects of GitHub’s Copilot on developer business leaders between April 2022 and May 2022.
productivity and happiness was sourced from the Thirteen countries were represented: Australia (100
GitHub Copilot Survey conducted in 2022. respondents), Brazil (115 respondents), Canada (175
The survey was emailed to 17,420 users who had opted respondents), China (200 respondents), France (130
in to receive communications and were using GitHub respondents), Germany (150 respondents), India
Copilot for their daily programming activities. Between (200 respondents), Israel (75 respondents), Japan
February 10, 2022, and March 6, 2022, the authors (100 respondents), Singapore (100 respondents),
received 2,047 responses that could be matched with South Africa (75 respondents), the United Kingdom
usage measurements during the four-week period (200 respondents), and the United States (1,000
leading up to March 12, 2022. The survey contained respondents). All participating companies have
multiple-choice questions on demographic information adopted AI technologies and are AI users.
and Likert-type questions on different aspects of Respondents were required to meet one of the
productivity, which were randomized in the order of following criteria: responsible for AI technology
appearance to the user. spending or approval of AI investments, developing
AI technology strategies, managing or overseeing
More details can be found in Ziegler at al., 2022.
AI technology implementation, serving as an AI
technology subject matter specialist, or making
Deloitte or influencing decisions around AI technology. To
Data used in the Corporate Activity-Industry Motivation complement the blind survey, Deloitte conducted
section was sourced from Deloitte’s “State of AI in the qualitative telephone interviews with 15 AI specialists
Enterprise” surveys. from various industries. More details are available on
Deloitte’s website.
More specifically, the AI Index made use of the following
sources of information:
International Federation of
Deloitte’s State of AI in the Enterprise,
5th Edition Report (2022)
Robotics (IFR)
State of AI in the Enterprise, 4th Edition (2021) Data presented in the Robot Installations section was

Deloitte’s State of AI in the Enterprise, 3rd Edition (2020) sourced from the “World Robotics 2022” report.

State of AI in the Enterprise, 2nd Edition (2018)

The 2017 Deloitte State of Cognitive Survey (2017)

Chapter 4 Preview 75