25 July 2025 | 12:10PM EDT

Global Economics Analyst

Quantifying the Risks of AI-Related Job Displacement

(Briggs/Dong)
n The recent pickup in AI adoption and reports of AI-related layoffs have raised Jan Hatzius
+1(212)902-0394 |
concerns that AI will lead to widespread labor displacement. In this Global Goldman Sachs & Co. LLC

Economics Analyst we examine the potential for AI-related labor disruptions. Joseph Briggs
+1(212)902-2163 |

n Although aggregate labor market impacts from generative AI remain limited, we Goldman Sachs & Co. LLC

are starting to see evidence of labor demand hits in the most AI-exposed Sarah Dong
+1(212)357-9741 |
industries. Employment growth has turned negative in marketing consulting, call Goldman Sachs & Co. LLC

centers, graphic design, web search and software development, and the tech Megan Peters
+44(20)7051-2058 |
sector’s employment share has declined below its long-run trend. Furthermore,
Goldman Sachs International
youth tech worker unemployment has risen by almost 3pp since the start of the
year, lending empirical support to anecdotes that AI is creating hiring headwinds
for recent tech graduates.
n While these trends could broaden as adoption increases, we remain skeptical
that AI will lead to large employment reductions over the next decade, primarily
because AI-related innovation will create new work opportunities that help offset
job losses from automation. That said, frictional unemployment typically
increases following adoption of labor-saving technologies, with each 1pp increase
in technology-driven productivity growth raising the unemployment rate by 0.3pp
in the near-term (but having no effect after two years).
n To assess the risk of temporary unemployment rate increases due to generative
AI, we develop a set of bottom-up displacement-risk metrics for over 800
occupations that predict whether AI productivity gains will translate to
displacement (rather than increases in output). Our baseline translation of our
risk measures to job loss suggests that AI will displace 6-7% of workers,
although displacement rates vary from 3-14% under different assumptions.
n Under our baseline displacement estimate and adoption timeline, we estimate
that generative AI will raise the unemployment rate by up to ½pp relative to its
trend during the AI transition period. We see risks as skewed towards a larger
increase in unemployment, particularly if AI adoption is more frontloaded than
we assume.

Investors should consider this report as only a single factor in making their investment decision. For Reg AC
certification and other important disclosures, see the Disclosure Appendix, or go to
[Link]/research/[Link].
Goldman Sachs Global Economics Analyst

Quantifying the Risks of AI-Related Job Displacement

We have extensively argued that generative artificial intelligence (AI) will provide
significant boosts to labor productivity and economic growth, primarily from its ability to
automate a large share of work tasks. While considerable uncertainty remains about
the timing and magnitude of AI’s effects, our baseline expectation remains that
generative AI will raise the level of labor productivity in the US and other DMs by around
15% following its full adoption and incorporation into regular production.

A recent pickup in AI adoption and reports of AI-related layoffs have raised concerns that
AI will lead to widespread labor displacement. Most notably, recent public company
commentary—particularly in tech and finance—indicates that executives see efficiency
gains from generative AI as a reason to slow hiring, especially for operational and
back-office workers (Appendix Exhibit 1). Examples include Amazon CEO Andrew Jassy
citing AI as a reason to reduce corporate headcount, JPMorgan consumer and
community banking CEO Marianne Lake suggesting that generative AI could reduce
headcount by 10% over the next five years, and Duolingo CEO Luis von Ahn and Shopify
CEO Tobi Lutke separately stating that teams must demonstrate that AI cannot
automate a position before hiring (our semiconductor’s equity analysts recently analyzed
the cost-saving implications of public commentary).

In this Global Economics Analyst, we examine the potential labor market disruption from
generative AI and quantify the impact on the unemployment rate under different
adoption and labor displacement scenarios.

Limited Macro but Emerging Micro Signs of AI’s Labor Market Impact
As we noted earlier this year, generative AI has had only a small impact on overall
macroeconomic and labor market statistics so far, primarily because adoption remains
low.

As shown in Exhibit 1, adoption rates have accelerated recently, particularly among

larger companies and in sectors with work tasks that are more exposed to AI-driven
automation. But the vast majority of companies still have not incorporated AI into
regular workflow, and only 9.3% of companies report that they have used generative AI
in production during the last two weeks.

25 July 2025 2
Goldman Sachs Global Economics Analyst

Exhibit 1: AI Adoption Has Accelerated Among Larger Companies and In More Exposed Industries, but Adoption Rates Remain Low

Source: Census Bureau, Goldman Sachs Global Investment Research

To illustrate the limited overall impact from AI, Exhibit 2 regresses our industry-level AI
exposure scores—which Exhibit 1 shows are highly correlated with adoption—with
different industry- and occupation-level labor market measures. This exercise yields no
statistically significant correlation between AI exposure and job growth, unemployment
rates, job finding rates, layoff rates, weekly hours growth, or average hourly earnings
growth.

Exhibit 2: AI-Exposure Is Not Correlated in an Economically or Statistically Meaningful Way With Most
Industry-Level Slack Measures

Coefficient Regressions of Industry-level Labor Coefficient on

on exposure Market Outcomes on AI Exposure exposure
0.20 0.20
0.15 Less slack
0.15
0.10 0.10
0.05 0.05
0.00 0.00
-0.05 -0.05
More slack
-0.10 -0.10
-0.15 -0.15
Job Growth Real Avg. Avg. Weekly Unemp. Rate Job Finding Layoff Rate
-0.20 -0.20
Hourly Hours Growth (inverted) Rate (inverted)
Earnings
Growth
Establishment Survey Household Survey

Source: Goldman Sachs Global Investment Research

Despite the limited economy-wide evidence, we do see early signs of labor market
disruption in specific industries where use cases have been established and anecdotal
reports suggest employment headwinds.

25 July 2025 3
Goldman Sachs Global Economics Analyst

As shown in the left chart of Exhibit 3, employment growth in industries like marketing
consulting, graphic design, office administration, and telephone call centers—all of
which have been reported as facing reduced labor demand due to AI-related efficiency
gains—has fallen well below trend. The right chart of Exhibit 3 shows that tech sector
employment growth has also slowed sharply.

Exhibit 3: Employment Is Contracting in Some Sectors Where Anecdotes Suggest AI Is Substituting for Labor

Note: Vertical lines are on November 2022, the month of ChatGPT’s public release. Trend is defined as average year-over-year growth rate from 2015-2019.

Source: US Bureau of Labor Statistics, Haver Analytics, Goldman Sachs Global Investment Research

The pullback in tech employment is especially notable when benchmarked against

longer-run trends. Exhibit 4 shows that tech employment (as a share of overall
employment) has decreased steadily since the public release of ChatGPT in November
2022. While some of the decline can be attributed to payback from over-hiring during
the pandemic, tech’s employment share has now fallen below its (remarkably linear)
pre-pandemic trend, a pattern probably related to AI automation.

25 July 2025 4
Goldman Sachs Global Economics Analyst

Exhibit 4: Tech’s Employment Share Has Declined Below Trend

Source: US Bureau of Labor Statistics, Haver Analytics, Goldman Sachs Global Investment Research

The recalibration of tech labor demand appears to be disproportionately affecting

younger tech workers whose skill set and capabilities more closely overlap with current
AI models. The left chart of Exhibit 5 shows that the unemployment rate for 20–
30-year-olds in tech-exposed occupations has risen by almost 3pp since the start of the
year, a notably larger increase than for either 20-30-year-olds employed in other
industries (+0.5pp) or overall tech workers (+1pp). This result lends empirical support to
anecdotal reports that generative AI is contributing to the hiring headwinds facing recent
college graduates in the tech sector.

Exhibit 5: Youth Unemployment in Tech Exposed Sectors Has Risen Recently

Note: Vertical line is on November 2022, the month of ChatGPT’s public release.

Source: IPUMS, Goldman Sachs Global Investment Research

The evidence that AI is already disrupting labor in specific industries (and tech in

25 July 2025 5
Goldman Sachs Global Economics Analyst

particular) is compelling. That said, the set of affected positions is quite narrow. To
quantify the near-term potential for labor disruption from current use cases, we collect
1) shares of total employment for occupations where AI has been anecdotally reported
to have already started replacing jobs and 2) the reported productivity boosts realized in
these occupations. Combining these data provides an upper bound on the share of total
employment currently “at risk” of automation.

The results of this exercise are shown in Exhibit 6. If current use cases were
extrapolated economy-wide and reduced employment by an amount equivalent to
reported productivity boosts, we estimate that 2.5% of US employment would be at
risk of displacement. While not economically insignificant, this low share highlights that
the vast majority of US labor is not at risk of automation today.

Exhibit 6: Early Evidence from Use Cases Where AI Is Already Driving Productivity Gains Suggests That At
Most 2.5% of Employment Is at Risk of Automation Today

Source: Data compiled by Goldman Sachs Global Investment Research, Haver Analytics, Goldman Sachs Global Investment Research

25 July 2025 6
Goldman Sachs Global Economics Analyst

Technology and Unemployment: Historical Evidence

Current data provides an incomplete picture of future automation risks given that AI
adoption will rise from its current 9% level in the coming years. We therefore turn to
historical evidence to assess the potential for labor market disruption as AI adoption
becomes more widespread.

There are two ways in which AI could lead to an increase in unemployment. First (and
the main focus of the most alarming commentary on AI-related job loss), technological
(or structural) unemployment could rise if AI capabilities advance to a point where
human input becomes redundant for many types of production. We are less concerned
about this outcome. Second (and more realistically), frictional unemployment could
increase in the near-term if it takes time for some workers that are displaced by AI to
find new jobs. We discuss both unemployment risks below.

Less Concern Regarding Persistent Technological Unemployment

The most sensational commentary regarding the unemployment risks from generative
AI largely focuses on the possibility that AI will eventually exceed human capabilities in
many areas, thereby leading to widespread and persistent job loss as human labor
becomes an inferior input to production. Most notably, Anthropic CEO Dario Amodei
has cited an upcoming “job apocalypse” where AI advances to a point where “cancer is
cured, the economy grows at 10% a year, the budget is balanced — and 20% of people
don’t have jobs.”

Predictions that technology will reduce the need for human labor have a long history but
a poor track record. For example, Nobel Prize-winning economist Wassily Leontief
famously predicted in 1983 that “the role of humans as the most important factor of
production is bound to diminish in the same way that the role of horses in agricultural
production was first diminished and then eliminated.” In particular, Leontief stated “I do
not see that the new industries can employ everybody who wants a job.’’

Despite predictions throughout history that technological advances would lead to

significant and persistent reductions in labor demand, employment has grown steadily.
The main reason that long-run technological unemployment has been averted is that,
while technology destroys jobs, it also raises demand for employees either directly in
new occupations that emerge due to technological change or indirectly following a
technology-driven increase in output and aggregate demand.

To demonstrate how technological innovation that initially displaces workers drives

employment growth over a long horizon, in Exhibit 7 we show results from a recent
study by economist David Autor and coauthors. Using Census data, they find that 60%
of workers today are employed in occupations that did not exist in 1940. Put differently,
over 85% of employment growth over the last 85 years is from the technology-driven
creation of new positions.

25 July 2025 7
Goldman Sachs Global Economics Analyst

Exhibit 7: Technological Innovation Leads to the Creation of New Occupations That Account for the Bulk of
Long-Run Employment Growth

Millions
Millions Employment,
Employment, byby New
New and
and Pre-Existing
Pre-Existing Occupations
Occupations Millions
Millions
200
200 Occupations That Existed 200 200
Occupations That Existed in in 1940
1940
175
175 Occupations
Occupations That
That Did
Did Not
Not Exist
Exist in in 1940
1940 175 175
150
150 150 150
125
125 125 125
100
100 100 100
75
75 75 75
50
50 50 50
25 25
25 25
0 0
0 0

Admin

Health

Total
Professionals

Production

Technicians
Managers

Sales
Transportation

Farming
Construction

Personal Services

Cleaning Services
Clerical&&Admin

Health

Total
Professionals

Transportation

Technicians
Managers

Sales
Production

Construction

Farming
Personal Services

Cleaning Services
Clerical

Source: Autor et al. (2022), Goldman Sachs Global Investment Research

A potential critique of this long-run evidence is that the advancements from generative
AI are different or could result in a level shift in productivity in ways that are not captured
by historical data. While it is hard to dismiss this possibility over the very long-run, the
well-documented limitations of current AI models as well as the continued human
outperformance in tasks involving the interaction between physical and cognitive
capabilities suggest that human labor will maintain a meaningful comparative advantage
in many areas of production for the foreseeable future.

The main risk to this view is that current model breakthroughs eventually lead to artificial
general intelligence (AGI), which we do not consider in our analysis. If this were to
occur, potential AI innovation could lead to an acceleration in productivity growth (rather
than the substantial but one-off increase in the level of productivity that boosts
productivity growth during the transition period that we assume) that eventually makes
human input in knowledge-based work tasks redundant. While this possibility has
received academic and media attention recently, a recent survey of experts suggested
that the average respondent saw only 50% odds that AI will be able to “accomplish
every task better and more cheaply than human workers” by 2047, with
“superforecasters” (i.e., individuals whose predictions have higher quality) putting only
50% odds of this occurring by 2081.

Provided that humans maintain competitive advantages in certain areas of production,

the AI-enabled increase in output and aggregate demand should continue to create new
employment opportunities for humans. As a result—and given that the historical
observation that humans are the only “low-cost, all-purpose computer system which
can be mass-produced by unskilled labor” still holds true for now—we remain skeptical
that AI productivity gains will result in long-run employment reductions over our 10-year
forecast horizon.

25 July 2025 8
Goldman Sachs Global Economics Analyst

Moderate Concern Regarding Frictional Employment

At the same time, the early evidence on hiring headwinds for specific industries
highlights the possibility that frictional unemployment could rise in the near-term if
AI-displaced workers experience an extended period of joblessness as they search for
new work.

Frictional unemployment is not unique to AI and occurs during most periods of rapid
technological change. To demonstrate this, in Exhibit 8 we estimate the historical
impulse response of the US unemployment rate to a 1pp increase in labor productivity
due to technological innovation, identified as in Gali 1999 by the assumption that
technology shocks lead to permanent increases in labor productivity while demand
shocks do not.

Our estimates suggest that a 1pp increase in labor productivity driven by technological
innovation leads to a temporary 0.3pp increase in the unemployment rate during the
first year after the shock but has no impact after two years (i.e., once displaced workers
find new jobs). Combining these estimates with our baseline forecast for a 15% uplift
to labor productivity from generative AI implies roughly 5pp of cumulative upward
pressure on unemployment from AI, or a roughly 0.3-0.5pp annual increase under a 10–
15-year adoption timeline.

Exhibit 8: Technology-Driven Productivity Gains Lead to Moderate but Temporary Upward Pressure on
Unemployment

Percentage Effect of a 1pp Increase in Labor Productivity Due to a Percentage

points Positive Technology Shock on the Unemployment Rate points
0.6 0.6

0.5 0.5

0.4 0.4

0.3 0.3

0.2 0.2

0.1 0.1

0 0

-0.1 -0.1

-0.2 -0.2
1 2 3 4 5 6 7 8
Quarters Since Shock

Source: Goldman Sachs Global Investment Research

This simple estimate is close to our long-standing assumption that workers with over
50% task exposure to AI are subject to automation, implying that 7% of workers will
ultimately be reallocated to new jobs due to AI in excess of current displacement
trends, which academic and OECD estimates suggest have historically subtracted
0.2-0.4% from employment growth per year (although these are offset by new job
creation; see our research on displacement trends during the last cycle). Our estimate
also aligns reasonably well with past periods of technological innovation, particularly the

25 July 2025 9
Goldman Sachs Global Economics Analyst

information technology (IT) automation boom.

An academic study by economists Daren Acemoglu and Pascual Restrepo decomposes

changes in the total wage bill into the contributions from technology-driven worker
displacement and reemployment and finds that automation of worker tasks lowered
labor demand by 0.3%/year (almost 10% cumulatively) from 1987-2017. Another study
by David Autor and Anna Salomons finds that from 1970-2007 automation directly
reduced labor demand in the US by 0.2%/year (8.2% cumulative). As noted in both
studies, overall labor demand continued to grow and displaced workers found new jobs,
largely due to increased aggregate demand.

Although it is hard to directly benchmark the productivity effects of generative AI against

IT, the historical evidence suggests that a 5-10% displacement rate and a less than ½pp
boost to the unemployment rate per year is a reasonable baseline for the frictional
unemployment impact from the productivity shock we envision from generative AI.

One of the biggest risks we see of a larger increase in the unemployment rate is that a
drop in aggregate demand could lead companies to accelerate plans to automate and
eliminate positions using generative AI if the economy enters a recession at some point
during the transition. As shown in Exhibit 9, historically nearly all of the job loss from
automation of routine occupations—defined as in Jaimovich Siu 2018—is concentrated
in economic downturns. If this dynamic repeats in the next few years, we could
imagine 1) more front-loaded job losses that push the unemployment rate to a much
higher level and 2) longer periods of joblessness if many similar workers are searching
for new jobs at the same time.

Exhibit 9: Labor Replacement Tends to Accelerate During Economic Downturns

Percent Workers Employed in Routine Occupations, Percent

40 Share of Population 40
38 38
36 36
34 34
32 32
30 30
28 28
26 26
24 24
22 22
20 20
1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 2018 2022

Source: Jaimovich Siu 2018, Goldman Sachs Global Investment Research

The main risk for a smaller increase in unemployment rate (even if the rate of
displacement increases) is that generative AI creates new profit pools and accelerates
the pace of innovation and new job creation. If so, this could provide an offsetting boost
to rehiring rates that limits the rise in frictional unemployment.

25 July 2025 10
Goldman Sachs Global Economics Analyst

Which Occupations Are More Likely to be Displaced?

Until the AI adoption cycle has fully played out, the potential labor market disruption—
including which jobs are likely to be displaced by generative AI—will remain an open
question.

To help assess this risk from a bottom-up perspective, we augment our long-standing AI
exposure scores with risk measures that quantify which occupations are more at risk of
displacement following the deployment of generative AI-enabled tools. In particular, we
consider five risk metrics based on corporate commentary and academic research
regarding which jobs have historically been displaced by automation:

n Consequence of Error – AI tools will never be completely failproof, and workers are
less likely to be displaced if mistakes expose companies to significant financial or
reputational risk. We measure consequences of errors via O*NET work context
scores that measure “How serious would the result usually be if the worker made a
mistake that was not easily correctable?” for each occupation.

n Task Repetitiveness – For a given level of AI task exposure, workers are more likely
to be displaced if their work involves more repetitive tasks that are easily
automated. We measure task repetitiveness via O*NET work context scores
regarding “How important are continuous, repetitive, physical activities (like key
entry) or mental activities (like checking entries in a ledger) to performing this job?”
for each occupation.

n Task Cohesion – As highlighted in Akçomak Borghans ter Weel 2011, automation has
historically led to more displacement in occupations with more connected tasks
(which raises the returns to automating multiple tasks at once). We measure task
cohesion (i.e., the extent to which work tasks belong together) based on the relative
frequency with which work tasks in O*NET are paired across occupations, with
higher scores indicating that a higher share of common or similar tasks that can be
more easily automated.

n Value of AI-Exposed Tasks vs. Overall Wage – Akçomak Borghans ter Weel 2011 also
show that displacement is less likely if automated tasks are lower value-add and in
areas where the worker does not have a comparative advantage. Following their
procedure, we proxy for comparative advantage in non-AI-exposed vs. AI-exposed
work activities as the difference between each occupation’s overall wage and the
wage rate for AI-exposed work tasks (estimated via a cross-sectional regression of
occupation-specific wages on occupation-specific time allocations).

n Back vs. Front Office – Corporate commentary has so far emphasized the potential
to automate back-office occupations. To classify all 832 occupations into back vs.
front office we 1) manually classify 80 occupations as back vs. front office, 2)
combine our manual classifications with measures of work contexts (e.g., frequency
of interaction with customers, importance of communications) to estimate a logistic
regression, and 3) use our estimated model to assign a “back-office” probability to
each occupation.

25 July 2025 11
Goldman Sachs Global Economics Analyst

Exhibits 10 and 11 show that our displacement risk metrics vary across occupations in
intuitive ways.

The left chart of Exhibit 10 shows replacement risk scores (normalized so that further
away from the center is associated with higher displacement risk) for seven occupations
that early evidence and anecdotes suggest are at higher risk of displacement. The right
chart shows the same for seven occupations that we believe to be less at risk for
displacement. For an interactive presentation of the risk scores for all occupations,
please see the appendix of the digital version of this report.

High-displacement risk occupations don’t necessarily score highly along all dimensions:
customer service representatives are generally front office positions, and the gap
between AI-exposed task wages and overall wages is small for legal secretaries and
accountants. Similarly, low-displacement risk occupations don’t score uniformly low on
all dimensions: radiologists’ and pharmacists’ AI-exposed tasks are fairly repetitive, and
the consequence of error for photographers is low. Nevertheless, the broad pattern
suggests higher displacement risk scores for occupations on the left and low scores on
the right, suggesting our displacement risk metrics are intuitively assessing
displacement risks correctly.

Exhibit 10: Our Displacement Risk Metrics Vary Across Occupations in an Intuitive Manner…
Highly at Risk Jobs Less at Risk Jobs

Computer Programmers Task Cohesion Air Traffic Controllers Task Cohesion

Accountants and Auditors Chief Executives
Legal/Administrative Assistants Radiologists
Customer Service Rep Pharmacists
Telemarketers Residential Advisors
Proofreaders & Copy Markers Photographers

AI-Task Wage Minus Back Office AI-Task Wage Minus Back Office
Overall Wage Probability Overall Wage Probability

Less
Less Risk
Risk

Task Decision
Task Decision
Repetiveness Consequence
Repetiveness Consequence

Note: All scores are expressed in standard units that are oriented such that a larger score represents greater risk of replacement. Scores exceeding +/-3 are winsorized to +/-3.

Source: O*NET, Goldman Sachs Global Investment Research

More systematically, Exhibit 11 averages across the five measures for each occupation
and presents the 15 occupations most at risk for displacement on the left and 15
occupations least at risk of displacement on the right. The patterns are again fairly
intuitive, with proofreaders, credit analysts, and telemarketers most at risk of
displacement as their AI-exposed tasks are automated, and dermatologists, door-to-door
salespeople, and preschool teachers least at risk of displacement.

25 July 2025 12
Goldman Sachs Global Economics Analyst

Exhibit 11: … And Predict Jobs That Are More or Less at Risk of Automation in a Manner That Aligns With Anecdotes

Source: O*NET, Goldman Sachs Global Investment Research

Despite these intuitive patterns, how the different considerations captured by our risk
metrics combine to determine which occupations are displaced will likely vary across
occupations in ways that are difficult to determine ahead of time. To provide insights
into potential displacement patterns, we conduct two exercises.

In Exhibit 12 we use the average of our five displacement risk scores and estimate how
the overall displacement rate would vary under various thresholds of occupational
displacement risk. We assume that employment in each occupation that is subject to
displacement is reduced by an amount equivalent to the productivity boost delivered by
full automation of AI-exposed tasks. We vary the threshold in 10pp increments.

In this exercise, we estimate that the economy-wide displacement rate for all workers
would total 2% if the 10% most at-risk occupations are ultimately displaced, 6% if the
top 20% are displaced, 9% if the top 30% are displaced, 11% if the top 40% are
displaced, and 14% if the top 50% are ultimately displaced. Our standing assumption
that productivity gains from generative AI ultimately will lead to the displacement of
6-7% of workers thus corresponds to a scenario where roughly the 20% most exposed
occupations experience employment headwinds from generative AI.

25 July 2025 13
Goldman Sachs Global Economics Analyst

Exhibit 12: Our Labor-at-Risk Scores Suggest a 2.4% Displacement Rate if the 10% Most At-Risk Jobs Are
Fully Automated, a 6.4% Displacement Rate if the Top 20% Are Fully Automated, and an 8.8% Displacement
Rate if the Top 30% Are Fully Automated

Source: Goldman Sachs Global Investment Research

Exhibit 13 examines a separate set of scenarios where the top 20% of occupations
most exposed to displacement according to each individual risk metric experience
employment reductions equivalent to our estimated AI-delivered productivity boosts. In
this scenario we estimate that economy-wide displacement rates would vary from 4%
to 9%, with smaller employment reductions if only back-office jobs are subject to
displacement, and larger employment reductions if the most repetitive tasks are subject
to displacement.

Exhibit 13: The Sectors Affected by Job Loss Vary Depending on the Importance of Different Components

Source: Goldman Sachs Global Investment Research

25 July 2025 14
Goldman Sachs Global Economics Analyst

Sizing the Macroeconomic Effects of AI-Related Job Displacement

Both the historical evidence and our bottom-up exercises point to AI-related job
displacement in the mid- to high-single digits, levels that are broadly in line with our
standing assumption of 7% cumulative labor displacement during the AI transition. To
estimate how displacement will translate to frictional unemployment, we combine this
assumption with our baseline adoption timeline, while also assuming that frictional
unemployment increases as productivity gains are realized but that most displaced
workers find new jobs after two years. We also translate the increases in economic
slack to an impact on inflation by applying a standard Phillips coefficient of 0.2.

The resulting scenarios are detailed in Exhibit 14. Under our baseline assumption, we
estimate that the unemployment rate rises roughly 1/2pp above trend in the first half of
the 2030s before fading in the second half of the decade. These estimates imply a
0.1pp drag on inflation that increases then fades along a similar time path from Phillips
curve channels.

While these outcomes are fairly benign, we see several risks around the AI outlook,
namely that AI adoption could be more frontloaded, that displacement rates could be
larger, or an economic downturn could lead to more front-loaded AI-related layoffs. We
therefore also consider scenarios where marginal adoption peaks in 2029, where the
40% most exposed workers (defined according to the average of our bottom-up risk
metrics) are subject to displacement, and where the first waves of AI-related
productivity gains primarily lead to displacement (until the cumulative displacement
threshold built into our estimates is exhausted). We also consider the interaction across
several of these scenarios.

Under these alternative assumptions, we estimate that the peak increase in the
unemployment rate could vary from 0.7-1.8pp (implying peak Phillips curve inflation
drags that vary from 0.15-0.4pp), and that the peak impacts could arrive as much as five
years earlier than in our baseline.

25 July 2025 15
Goldman Sachs Global Economics Analyst

Exhibit 14: Our Baseline Estimates Imply a 0.4pp Peak Increase in the Unemployment Rate and 0.1pp Drag
on Inflation Due to AI Automation, but See Risks Skewed Toward Larger and Earlier Effects

Percentage Effect of AI-Driven Labor Displacement on the Percentage points

Unemployment Rate by Scenario
2.0 2.0
Baseline
1.8 1.8
Faster Adoption
1.6 Larger Displacement 1.6
1.4 Frontloaded Displacement 1.4
Faster x Larger
1.2 Frontloaded x Larger 1.2
1.0 1.0
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0.0 0.0
2020 2025 2030 2035 2040 2045 2050

Source: Goldman Sachs Global Investment Research

Taken together, our analysis suggests that labor displacement from generative AI will
have a meaningful but manageable impact on the US economy over the next 5-10 years.
That said, risks are skewed toward a more disruptive impact, particularly if AI adoption is
more frontloaded than we expect.

Joseph Briggs

Sarah Dong

25 July 2025 16
Goldman Sachs Global Economics Analyst

Appendix

Appendix Exhibit 1 – Recent Anecdotes and Company Commentary Suggest Companies Are Starting to Reduce Headcount Due to
AI-Generated Efficiencies
Company Speaker Quote Date

AI is doing 30 to 50% of the work at Salesforce now and I think that that will, you know,
Marc Benioff
Salesforce continue. I think that all of us have to get our head around this idea that AI can do things that 6/26/2025
(CEO)
before, you know, we were doing and we can move on to doing higher value work.

As we roll out more Generative AI and agents, it should change the way our work is done.
We will need fewer people doing some of the jobs that are being done today, and more
Andy Jassy
Amazon people doing other types of jobs. It’s hard to know exactly where this nets out over time, but 6/17/2025
(CEO)
in the next few years, we expect that this will reduce our total corporate workforce as we get
efficiency gains from using AI extensively across the company.

The operations team is at the tip of the spear on using and leveraging new AI tools and
Marianne Lake
capabilities. And based upon what we know today, we expect headcount will trend down by
JPMorgan Chase (CEO of Consumer & 5/19/2025
about 10% over the next five years or so, even as the business grows by another more than
Community Banking)
25%.

When we think about [the] HR industry, which is [a] $300 billion-plus industry but its including
Recruit Holdings
Deko Idekoba 60 or 65% of human labor manual cost. It's very difficult to find [an industry] with such a high
(parent of Indeed and 5/14/2025
(CEO) percentage of human labor manual cost. What we believe is basically, how can we simplify
Glassdoor)
hiring with using AI and technology and data to reduce manual work?

Duolingo is going to be AI-first. AI is already changing how work gets done. It's not a
question or if or when. It's happening now. … Being AI-first means we will need to rethink
much of how we work.

Luis von Ahn We'll be rolling out a few constructive constraints to help guide this:
Duolingo 4/28/2025
(CEO) - We'll gradually stop using contractors to do work that AI can handle
- AI use will be part of what we look for in hiring
- AI use will be part of what we evaluate in performance reviews
- Headcount will only be given if a team cannot automate more of their work
- Most functions will have specific initiatives to fundamentally change how they work

So here is the unpleasant truth: AI is coming for your jobs. Heck, it's coming for my job too.
Micha Kaufman This is a wake-up call. It does not matter if you are a programmer, designer, product
Fiverr 4/7/2025
(CEO) manager, data scientist, lawyer, customer support rep, salesperson, or a finance person - AI
is coming for you.

Before asking for more [h]eadcount and resources, teams must demonstrate why they
Tobi Lutke cannot get what they want done using AI. What would this area look like if autonomous AI
Shopify 4/7/2025
(CEO) agents were already part of the team? This question can lead to really fun discussions and
projects.

Source: Goldman Sachs Global Investment Research

25 July 2025 17
Goldman Sachs Global Economics Analyst

Disclosure Appendix
Reg AC
We, Jan Hatzius, Joseph Briggs, Sarah Dong and Megan Peters, hereby certify that all of the views expressed in this report accurately reflect our
personal views, which have not been influenced by considerations of the firm’s business or client relationships.
Unless otherwise stated, the individuals listed on the cover page of this report are analysts in Goldman Sachs’ Global Investment Research division.

Disclosures
Regulatory disclosures
Disclosures required by United States laws and regulations
See company-specific regulatory disclosures above for any of the following disclosures required as to companies referred to in this report: manager or
co-manager in a pending transaction; 1% or other ownership; compensation for certain services; types of client relationships; managed/co-managed
public offerings in prior periods; directorships; for equity securities, market making and/or specialist role. Goldman Sachs trades or may trade as a
principal in debt securities (or in related derivatives) of issuers discussed in this report.
The following are additional required disclosures: Ownership and material conflicts of interest: Goldman Sachs policy prohibits its analysts,
professionals reporting to analysts and members of their households from owning securities of any company in the analyst’s area of coverage.
Analyst compensation: Analysts are paid in part based on the profitability of Goldman Sachs, which includes investment banking revenues. Analyst
as officer or director: Goldman Sachs policy generally prohibits its analysts, persons reporting to analysts or members of their households from
serving as an officer, director or advisor of any company in the analyst’s area of coverage. Non-U.S. Analysts: Non-U.S. analysts may not be
associated persons of Goldman Sachs & Co. LLC and therefore may not be subject to FINRA Rule 2241 or FINRA Rule 2242 restrictions on
communications with subject company, public appearances and trading securities held by the analysts.

Additional disclosures required under the laws and regulations of jurisdictions other than the United States
The following disclosures are those required by the jurisdiction indicated, except to the extent already made above pursuant to United States laws and
regulations. Australia: Goldman Sachs Australia Pty Ltd and its affiliates are not authorised deposit-taking institutions (as that term is defined in the
Banking Act 1959 (Cth)) in Australia and do not provide banking services, nor carry on a banking business, in Australia. This research, and any access to
it, is intended only for “wholesale clients” within the meaning of the Australian Corporations Act, unless otherwise agreed by Goldman Sachs. In
producing research reports, members of Global Investment Research of Goldman Sachs Australia may attend site visits and other meetings hosted by
the companies and other entities which are the subject of its research reports. In some instances the costs of such site visits or meetings may be met
in part or in whole by the issuers concerned if Goldman Sachs Australia considers it is appropriate and reasonable in the specific circumstances relating
to the site visit or meeting. To the extent that the contents of this document contains any financial product advice, it is general advice only and has
been prepared by Goldman Sachs without taking into account a client’s objectives, financial situation or needs. A client should, before acting on any
such advice, consider the appropriateness of the advice having regard to the client’s own objectives, financial situation and needs. A copy of certain
Goldman Sachs Australia and New Zealand disclosure of interests and a copy of Goldman Sachs’ Australian Sell-Side Research Independence Policy
Statement are available at: [Link] Brazil: Disclosure information in relation to
CVM Resolution n. 20 is available at [Link] Where applicable, the Brazil-registered analyst primarily
responsible for the content of this research report, as defined in Article 20 of CVM Resolution n. 20, is the first author named at the beginning of this
report, unless indicated otherwise at the end of the text. Canada: This information is being provided to you for information purposes only and is not,
and under no circumstances should be construed as, an advertisement, offering or solicitation by Goldman Sachs & Co. LLC for purchasers of
securities in Canada to trade in any Canadian security. Goldman Sachs & Co. LLC is not registered as a dealer in any jurisdiction in Canada under
applicable Canadian securities laws and generally is not permitted to trade in Canadian securities and may be prohibited from selling certain securities
and products in certain jurisdictions in Canada. If you wish to trade in any Canadian securities or other products in Canada please contact Goldman
Sachs Canada Inc., an affiliate of The Goldman Sachs Group Inc., or another registered Canadian dealer. Hong Kong: Further information on the
securities of covered companies referred to in this research may be obtained on request from Goldman Sachs (Asia) L.L.C. India: Further information
on the subject company or companies referred to in this research may be obtained from Goldman Sachs (India) Securities Private Limited, Research
Analyst - SEBI Registration Number INH000001493, 951-A, Rational House, Appasaheb Marathe Marg, Prabhadevi, Mumbai 400 025, India, Corporate
Identity Number U74140MH2006FTC160634, Phone +91 22 6616 9000, Fax +91 22 6616 9001. Goldman Sachs may beneficially own 1% or more of
the securities (as such term is defined in clause 2 (h) the Indian Securities Contracts (Regulation) Act, 1956) of the subject company or companies
referred to in this research report. Investment in securities market are subject to market risks. Read all the related documents carefully before
investing. Registration granted by SEBI and certification from NISM in no way guarantee performance of the intermediary or provide any assurance of
returns to investors. Goldman Sachs (India) Securities Private Limited compliance officer and investor grievance contact details can be found at this link:
[Link] - /general/equity. Japan: See below. Korea: This research, and any access to it, is intended only for
“professional investors” within the meaning of the Financial Services and Capital Markets Act, unless otherwise agreed by Goldman Sachs. Further
information on the subject company or companies referred to in this research may be obtained from Goldman Sachs (Asia) L.L.C., Seoul Branch. New
Zealand: Goldman Sachs New Zealand Limited and its affiliates are neither “registered banks” nor “deposit takers” (as defined in the Reserve Bank of
New Zealand Act 1989) in New Zealand. This research, and any access to it, is intended for “wholesale clients” (as defined in the Financial Advisers Act
2008) unless otherwise agreed by Goldman Sachs. A copy of certain Goldman Sachs Australia and New Zealand disclosure of interests is available at:
[Link] Russia: Research reports distributed in the Russian Federation are not
advertising as defined in the Russian legislation, but are information and analysis not having product promotion as their main purpose and do not
provide appraisal within the meaning of the Russian legislation on appraisal activity. Research reports do not constitute a personalized investment
recommendation as defined in Russian laws and regulations, are not addressed to a specific client, and are prepared without analyzing the financial
circumstances, investment profiles or risk profiles of clients. Goldman Sachs assumes no responsibility for any investment decisions that may be taken
by a client or any other person based on this research report. Singapore: Goldman Sachs (Singapore) Pte. (Company Number: 198602165W), which is
regulated by the Monetary Authority of Singapore, accepts legal responsibility for this research, and should be contacted with respect to any matters
arising from, or in connection with, this research. Taiwan: This material is for reference only and must not be reprinted without permission. Investors
should carefully consider their own investment risk. Investment results are the responsibility of the individual investor. United Kingdom: Persons who
would be categorized as retail clients in the United Kingdom, as such term is defined in the rules of the Financial Conduct Authority, should read this
research in conjunction with prior Goldman Sachs research on the covered companies referred to herein and should refer to the risk warnings that have
been sent to them by Goldman Sachs International. A copy of these risks warnings, and a glossary of certain financial terms used in this report, are
available from Goldman Sachs International on request.
European Union and United Kingdom: Disclosure information in relation to Article 6 (2) of the European Commission Delegated Regulation (EU)
(2016/958) supplementing Regulation (EU) No 596/2014 of the European Parliament and of the Council (including as that Delegated Regulation is

25 July 2025 18
Goldman Sachs Global Economics Analyst

implemented into United Kingdom domestic law and regulation following the United Kingdom’s departure from the European Union and the European
Economic Area) with regard to regulatory technical standards for the technical arrangements for objective presentation of investment
recommendations or other information recommending or suggesting an investment strategy and for disclosure of particular interests or indications of
conflicts of interest is available at [Link] which states the European Policy for Managing Conflicts of
Interest in Connection with Investment Research.
Japan: Goldman Sachs Japan Co., Ltd. is a Financial Instrument Dealer registered with the Kanto Financial Bureau under registration number Kinsho
69, and a member of Japan Securities Dealers Association, Financial Futures Association of Japan Type II Financial Instruments Firms Association, The
Investment Trusts Association, Japan, and Japan Investment Advisers Association. Sales and purchase of equities are subject to commission
pre-determined with clients plus consumption tax. See company-specific disclosures as to any applicable disclosures required by Japanese stock
exchanges, the Japanese Securities Dealers Association or the Japanese Securities Finance Company.

Global product; distributing entities

Goldman Sachs Global Investment Research produces and distributes research products for clients of Goldman Sachs on a global basis. Analysts based
in Goldman Sachs offices around the world produce research on industries and companies, and research on macroeconomics, currencies, commodities
and portfolio strategy. This research is disseminated in Australia by Goldman Sachs Australia Pty Ltd (ABN 21 006 797 897); in Brazil by Goldman Sachs
do Brasil Corretora de Títulos e Valores Mobiliários S.A.; Public Communication Channel Goldman Sachs Brazil: 0800 727 5764 and / or
contatogoldmanbrasil@[Link]. Available Weekdays (except holidays), from 9am to 6pm. Canal de Comunicação com o Público Goldman Sachs Brasil:
0800 727 5764 e/ou contatogoldmanbrasil@[Link]. Horário de funcionamento: segunda-feira à sexta-feira (exceto feriados), das 9h às 18h; in Canada
by Goldman Sachs & Co. LLC; in Hong Kong by Goldman Sachs (Asia) L.L.C.; in India by Goldman Sachs (India) Securities Private Ltd.; in Japan by
Goldman Sachs Japan Co., Ltd.; in the Republic of Korea by Goldman Sachs (Asia) L.L.C., Seoul Branch; in New Zealand by Goldman Sachs New
Zealand Limited; in Russia by OOO Goldman Sachs; in Singapore by Goldman Sachs (Singapore) Pte. (Company Number: 198602165W); and in the
United States of America by Goldman Sachs & Co. LLC. Goldman Sachs International has approved this research in connection with its distribution in
the United Kingdom.
Goldman Sachs International (“GSI”), authorised by the Prudential Regulation Authority (“PRA”) and regulated by the Financial Conduct Authority
(“FCA”) and the PRA, has approved this research in connection with its distribution in the United Kingdom.
European Economic Area: GSI, authorised by the PRA and regulated by the FCA and the PRA, disseminates research in the following jurisdictions
within the European Economic Area: the Grand Duchy of Luxembourg, Italy, the Kingdom of Belgium, the Kingdom of Denmark, the Kingdom of
Norway, the Republic of Finland and the Republic of Ireland; GSI - Succursale de Paris (Paris branch) which is authorised by the French Autorité de
contrôle prudentiel et de resolution (“ACPR”) and regulated by the Autorité de contrôle prudentiel et de resolution and the Autorité des marches
financiers (“AMF”) disseminates research in France; GSI - Sucursal en España (Madrid branch) authorized in Spain by the Comisión Nacional del
Mercado de Valores disseminates research in the Kingdom of Spain; GSI - Sweden Bankfilial (Stockholm branch) is authorized by the SFSA as a “third
country branch” in accordance with Chapter 4, Section 4 of the Swedish Securities and Market Act (Sw. lag (2007:528) om värdepappersmarknaden)
disseminates research in the Kingdom of Sweden; Goldman Sachs Bank Europe SE (“GSBE”) is a credit institution incorporated in Germany and, within
the Single Supervisory Mechanism, subject to direct prudential supervision by the European Central Bank and in other respects supervised by German
Federal Financial Supervisory Authority (Bundesanstalt für Finanzdienstleistungsaufsicht, BaFin) and Deutsche Bundesbank and disseminates research
in the Federal Republic of Germany and those jurisdictions within the European Economic Area where GSI is not authorised to disseminate research
and additionally, GSBE, Copenhagen Branch filial af GSBE, Tyskland, supervised by the Danish Financial Authority disseminates research in the Kingdom
of Denmark; GSBE - Sucursal en España (Madrid branch) subject (to a limited extent) to local supervision by the Bank of Spain disseminates research in
the Kingdom of Spain; GSBE - Succursale Italia (Milan branch) to the relevant applicable extent, subject to local supervision by the Bank of Italy (Banca
d’Italia) and the Italian Companies and Exchange Commission (Commissione Nazionale per le Società e la Borsa “Consob”) disseminates research in
Italy; GSBE - Succursale de Paris (Paris branch), supervised by the AMF and by the ACPR disseminates research in France; and GSBE - Sweden
Bankfilial (Stockholm branch), to a limited extent, subject to local supervision by the Swedish Financial Supervisory Authority (Finansinpektionen)
disseminates research in the Kingdom of Sweden.

General disclosures
This research is for our clients only. Other than disclosures relating to Goldman Sachs, this research is based on current public information that we
consider reliable, but we do not represent it is accurate or complete, and it should not be relied on as such. The information, opinions, estimates and
forecasts contained herein are as of the date hereof and are subject to change without prior notification. We seek to update our research as
appropriate, but various regulations may prevent us from doing so. Other than certain industry reports published on a periodic basis, the large majority
of reports are published at irregular intervals as appropriate in the analyst’s judgment.
Goldman Sachs conducts a global full-service, integrated investment banking, investment management, and brokerage business. We have investment
banking and other business relationships with a substantial percentage of the companies covered by Global Investment Research. Goldman Sachs &
Co. LLC, the United States broker dealer, is a member of SIPC ([Link]
Our salespeople, traders, and other professionals may provide oral or written market commentary or trading strategies to our clients and principal
trading desks that reflect opinions that are contrary to the opinions expressed in this research. Our asset management area, principal trading desks and
investing businesses may make investment decisions that are inconsistent with the recommendations or views expressed in this research.
We and our affiliates, officers, directors, and employees will from time to time have long or short positions in, act as principal in, and buy or sell, the
securities or derivatives, if any, referred to in this research, unless otherwise prohibited by regulation or Goldman Sachs policy.
The views attributed to third party presenters at Goldman Sachs arranged conferences, including individuals from other parts of Goldman Sachs, do not
necessarily reflect those of Global Investment Research and are not an official view of Goldman Sachs.
Any third party referenced herein, including any salespeople, traders and other professionals or members of their household, may have positions in the
products mentioned that are inconsistent with the views expressed by analysts named in this report.
This research is focused on investment themes across markets, industries and sectors. It does not attempt to distinguish between the prospects or
performance of, or provide analysis of, individual companies within any industry or sector we describe.
Any trading recommendation in this research relating to an equity or credit security or securities within an industry or sector is reflective of the
investment theme being discussed and is not a recommendation of any such security in isolation.
This research is not an offer to sell or the solicitation of an offer to buy any security in any jurisdiction where such an offer or solicitation would be
illegal. It does not constitute a personal recommendation or take into account the particular investment objectives, financial situations, or needs of
individual clients. Clients should consider whether any advice or recommendation in this research is suitable for their particular circumstances and, if
appropriate, seek professional advice, including tax advice. The price and value of investments referred to in this research and the income from them
may fluctuate. Past performance is not a guide to future performance, future returns are not guaranteed, and a loss of original capital may occur.
Fluctuations in exchange rates could have adverse effects on the value or price of, or income derived from, certain investments.

25 July 2025 19
Goldman Sachs Global Economics Analyst

Certain transactions, including those involving futures, options, and other derivatives, give rise to substantial risk and are not suitable for all investors.
Investors should review current options and futures disclosure documents which are available from Goldman Sachs sales representatives or at
[Link] and [Link] Transaction costs
may be significant in option strategies calling for multiple purchase and sales of options such as spreads. Supporting documentation will be supplied
upon request.
Differing Levels of Service provided by Global Investment Research: The level and types of services provided to you by Goldman Sachs Global
Investment Research may vary as compared to that provided to internal and other external clients of GS, depending on various factors including your
individual preferences as to the frequency and manner of receiving communication, your risk profile and investment focus and perspective (e.g.,
marketwide, sector specific, long term, short term), the size and scope of your overall client relationship with GS, and legal and regulatory constraints.
As an example, certain clients may request to receive notifications when research on specific securities is published, and certain clients may request
that specific data underlying analysts’ fundamental analysis available on our internal client websites be delivered to them electronically through data
feeds or otherwise. No change to an analyst’s fundamental research views (e.g., ratings, price targets, or material changes to earnings estimates for
equity securities), will be communicated to any client prior to inclusion of such information in a research report broadly disseminated through electronic
publication to our internal client websites or through other means, as necessary, to all clients who are entitled to receive such reports.
All research reports are disseminated and available to all clients simultaneously through electronic publication to our internal client websites. Not all
research content is redistributed to our clients or available to third-party aggregators, nor is Goldman Sachs responsible for the redistribution of our
research by third party aggregators. For research, models or other data related to one or more securities, markets or asset classes (including related
services) that may be available to you, please contact your GS representative or go to [Link]
Disclosure information is also available at [Link] or from Research Compliance, 200 West Street, New York, NY
10282.
© 2025 Goldman Sachs.
You are permitted to store, display, analyze, modify, reformat, and print the information made available to you via this service only for your own use.
You may not resell or reverse engineer this information to calculate or develop any index for disclosure and/or marketing or create any other derivative
works or commercial product(s), data or offering(s) without the express written consent of Goldman Sachs. You are not permitted to publish, transmit,
or otherwise reproduce this information, in whole or in part, in any format to any third party without the express written consent of Goldman Sachs.
This foregoing restriction includes, without limitation, using, extracting, downloading or retrieving this information, in whole or in part, to train or
finetune a machine learning or artificial intelligence system, or to provide or reproduce this information, in whole or in part, as a prompt or input to any
such system.

25 July 2025 20