🤖
AI's Impact on the Labor Market: A New Measurement and Early Evidence
Economics
This report introduces a novel measure for assessing AI's impact on the labor market, combining theoretical LLM capabilities with real-world usage data. It analyzes early evidence, finding that while AI is far from reaching its full potential, occupations with higher observed exposure are projected to grow less. The study also notes that workers in more exposed professions tend to be older, female, more educated, and higher-paid, but currently observes no systematic increase in unemployment. However, there is suggestive evidence of a slowdown in hiring for younger workers in exposed occupations.
Artificial Intelligence
Labor Economics
Future of Work
29 Questions
Medium
Ages 16+
Mar 9, 2026
About this Study Set
This study set covers Economics through
29 practice questions.
This report introduces a novel measure for assessing AI's impact on the labor market, combining theoretical LLM capabilities with real-world usage data. It analyzes early evidence, finding that while AI is far from reaching its full potential, occupations with higher observed exposure are projected to grow less. The study also notes that workers in more exposed professions tend to be older, female, more educated, and higher-paid, but currently observes no systematic increase in unemployment. However, there is suggestive evidence of a slowdown in hiring for younger workers in exposed occupations. Every question includes the correct answer so you can learn as you go — pick any format above to get started.
Questions & Answers
Browse all 29 questions from the
AI's Impact on the Labor Market: A New Measurement and Early Evidence study set below.
Each question shows the correct answer — select a study format above to practice interactively.
1
What is the primary goal of the new measure introduced in this paper?
-
A
To predict future AI capabilities
-
B
To measure AI displacement risk and observed exposure
-
C
To assess the theoretical limits of LLMs
-
D
To track the historical adoption of AI
2
Which of the following is NOT a component of the new AI displacement risk measure, Observed Exposure?
-
A
Theoretical LLM capability
-
B
Real-world usage data
-
C
Worker demographics
-
D
Weighting of automated uses
3
According to the findings, how does actual AI coverage compare to its theoretical capability?
-
A
Actual coverage exceeds theoretical capability
-
B
Actual coverage remains a fraction of what's feasible
-
C
Actual coverage is equal to theoretical capability
-
D
Actual coverage is not measured
4
Which government agency's projections are used to assess the growth of occupations with higher observed AI exposure?
-
A
Bureau of Labor Statistics (BLS)
-
B
National Science Foundation (NSF)
-
C
Department of Commerce
-
D
Federal Reserve
5
What demographic characteristics are more likely to be found in workers in the most exposed professions?
-
A
Younger, male, less educated, lower-paid
-
B
Older, female, more educated, higher-paid
-
C
Younger, female, less educated, lower-paid
-
D
Older, male, more educated, higher-paid
6
What has been the observed impact on unemployment for highly exposed workers since late 2022?
-
A
A significant increase in unemployment
-
B
A slight decrease in unemployment
-
C
No systematic increase in unemployment
-
D
Unemployment data is not yet available
7
What suggestive evidence regarding younger workers has been found in exposed occupations?
-
A
Hiring has increased
-
B
Separations have increased
-
C
Hiring has slowed
-
D
Unemployment has significantly increased
8
What is one of the key challenges in measuring the labor market impacts of AI, as mentioned in the text?
-
A
AI's impact is always immediately obvious
-
B
Past approaches have a strong track record of prediction
-
C
The effects may be subtle and clouded by other economic factors
-
D
AI's impact is uniform across all industries
9
The paper compares the potential impact of AI to which historical economic disruptions?
-
A
The Industrial Revolution and the Great Depression
-
B
The COVID-19 pandemic and the internet
-
C
The invention of the printing press and the steam engine
-
D
The agricultural revolution and the invention of the wheel
10
What is the primary goal of the new framework presented in the paper?
-
A
To immediately replace human workers with AI
-
B
To establish an approach for measuring AI's employment effects and revisit analyses periodically
-
C
To prove AI's superiority over human labor
-
D
To forecast AI's impact on job creation only
11
When are AI's impacts most likely to be unmistakable and easily identifiable?
-
A
When they are subtle and gradual
-
B
When they are large and sudden, like COVID-19
-
C
When they only affect theoretical capabilities
-
D
When they are limited to augmentative uses
12
How is 'exposure' typically defined at the task level in AI impact studies?
-
A
By the theoretical possibility of AI performing the task
-
B
By the number of workers who use AI tools
-
C
By the cost of AI implementation
-
D
By the speed at which AI can complete a task
13
Which database is used to enumerate tasks associated with US occupations?
-
A
LinkedIn
-
B
O*NET
-
C
Indeed
-
D
Glassdoor
14
What is the 'Anthropic Economic Index' used for in this study?
-
A
To measure theoretical LLM capabilities
-
B
To measure real-world usage data of AI
-
C
To forecast future job growth
-
D
To track worker demographics
15
According to Eloundou et al. (2023), what does a β score of 1 indicate for a task?
-
A
The task is not feasible for an LLM
-
B
The task can be doubled in speed by an LLM alone
-
C
The task requires additional software to be sped up by an LLM
-
D
The task is only partially feasible for an LLM
16
What is a potential reason why actual AI usage might fall short of theoretical capability?
-
A
LLMs are always perfect and have no limitations
-
B
Legal constraints, software requirements, or human verification steps
-
C
Lack of interest from businesses
-
D
AI is too expensive for widespread adoption
17
What does the 'Observed Exposure' measure quantify?
-
A
The theoretical maximum tasks an LLM could perform
-
B
The gap between theoretical capability and actual automated usage in professional settings
-
C
The number of AI developers worldwide
-
D
The total investment in AI research
18
Which of the following is NOT listed as a factor that increases a job's 'Observed Exposure'?
-
A
Tasks are theoretically possible with AI
-
B
Tasks see significant usage in the Anthropic Economic Index
-
C
Tasks are performed in non-work-related contexts
-
D
Tasks have a relatively higher share of automated use patterns
19
How are fully automated implementations weighted in the 'Observed Exposure' calculation compared to augmentative use?
-
A
Fully automated implementations receive half weight
-
B
Fully automated implementations receive full weight
-
C
They receive equal weight
-
D
They are not weighted differently
20
Figure 2 shows that AI is far from reaching its theoretical capabilities. What percentage of tasks in the Computer & Math category does Claude currently cover?
21
According to Figure 3, which occupation is listed as the most exposed with 75% coverage?
-
A
Customer Service Representatives
-
B
Data Entry Keyers
-
C
Computer Programmers
-
D
Financial Analysts
22
Which of the following occupations is NOT listed as having zero coverage due to infrequent appearance in the data?
-
A
Cooks
-
B
Motorcycle Mechanics
-
C
Lifeguards
-
D
Baristas
23
A regression analysis found that for every 10 percentage point increase in observed exposure, BLS growth projections dropped by how much?
-
A
1.0 percentage point
-
B
0.6 percentage points
-
C
0.1 percentage points
-
D
5.0 percentage points
24
In the three months before ChatGPT's release, workers in the most exposed professions were more likely to be:
-
A
Younger and male
-
B
Older and female
-
C
Younger and female
-
D
Older and male
25
What is the primary outcome the researchers focus on to capture potential economic harm?
-
A
Changes in job postings
-
B
Job creation in new sectors
-
C
Increased unemployment
-
D
Productivity gains
26
Figure 6 shows trends in the unemployment rate. What is observed about the gap between the most and least AI-exposed workers since ChatGPT's release?
-
A
The gap has significantly widened
-
B
The gap has remained consistently large
-
C
The average change in the gap is small and insignificant
-
D
The gap has closed entirely
27
What concern is raised regarding young workers and AI-exposed occupations?
-
A
A significant increase in their unemployment rate
-
B
A slowdown in their hiring into these occupations
-
C
A decrease in their education levels
-
D
An increase in their job separations
28
Figure 7 shows the job start rate for young workers. What trend is observed for entry into high-exposure occupations after ChatGPT's era?
-
A
It has remained stable
-
B
It has increased by half a percentage point
-
C
It has decreased by about half a percentage point
-
D
It has become more volatile
29
What is a potential alternative interpretation for young workers not being hired into exposed occupations?
-
A
They are all choosing to become unemployed
-
B
They are remaining at existing jobs, taking different jobs, or returning to school
-
C
AI is directly causing them to quit their current jobs
-
D
They are unable to find any jobs at all
