Labour Market Impacts of Artificial Intelligence

A recent research report by Anthropic, analysing AI usage data and labour statistics from the S. Bureau of Labor Statistics, introduces a new framework called “observed exposure” to measure how AI is affecting the labour market in practice rather than merely in theory.

Table of Contents

Context

A recent research report by Anthropic, analysing AI usage data and labour statistics from the S. Bureau of Labor Statistics, introduces a new framework called “observed exposure” to measure how AI is affecting the labour market in practice rather than merely in theory.

A New Approach: Concept of “Observed Exposure”

● Observed exposure is a new indicator that measures the real-world interaction between AI systems and occupational tasks. It combines three key elements:

● Theoretical AI capability — whether AI could perform a task

● Actual usage data from AI platforms

● Nature of usage — whether AI automates or merely assists a task

● The measure emphasises automation-related usage rather than augmentation, meaning tasks fully executed by AI carry greater weight.

● This approach helps bridge the gap between what AI could theoretically do and what it is actually doing in workplaces today.

Key observations

Gap Between AI Capability and Actual Usage: The study finds that AI usage remains far below its theoretical capability.
- g. In Computer and Mathematical occupations, AI could theoretically assist in 94% of tasks, but current usage covers only about 33%.

This gap exists because:

AI tools are still evolving
Regulatory and legal constraints slow adoption
Human verification is required
Organisational inertia delays technological diffusion

Tasks Beyond AI’s Reach: Many tasks remain difficult or impossible for AI due to physical or legal constraints. Examples include:
- Agricultural work such as pruning trees
- Skilled mechanical repair
- Courtroom legal representation

These tasks involve physical manipulation, human judgment, or regulatory responsibility, limiting automation.

Occupations Most Exposed to AI: The study identifies several professions where AI usage already covers a large portion of tasks.
- Computer Programmers: Coding-related tasks are highly compatible with AI assistance, with coverage estimated at around 75%.
- Customer Service Representatives: AI chatbots and automated response systems are increasingly performing customer support functions.
- Data Entry Workers: Routine tasks such as entering structured data into systems are easily automated by AI tools.
- Financial Analysts: AI is widely used for data analysis, summarisation, and report generation in finance.
Occupations Least Exposed to AI: Around 30% of workers have zero measurable exposure to AI automation. Examples include Cook, Mechanics, Lifeguards, Bartenders and Dishwashers
- These jobs involve physical labour and interpersonal interaction, which remain difficult for current AI systems.

Demographic Characteristics of Highly Exposed Workers

The research finds that workers in AI-exposed occupations have distinct characteristics. Compared to workers in non-exposed jobs, they are More educated, Higher paid, More likely to be female and More likely to work in white-collar professions
- g. Workers with graduate degrees make up 17.4% of the most exposed group, compared to 4.5% among unexposed workers.
This suggests that AI may primarily affect white-collar knowledge work rather than manual labour.

Impact on Employment and Job Growth

Relationship with Future Employment Growth: Using projections from the S. Bureau of Labor Statistics, the study finds that occupations with higher AI exposure are expected to grow slightly more slowly.
- g. Every 10-percentage-point increase in AI exposure reduces projected job growth by about 0.6 percentage points.
No Significant Rise in Unemployment: Despite rapid AI adoption since 2022, the study finds there is no systematic increase in unemployment in highly exposed occupations.
- Unemployment trends in exposed and non-exposed occupations have remained broadly similar. This suggests that AI is currently augmenting work rather than replacing workers.
Hiring Slowdown for Young Workers: The research identifies one emerging labour-market signal.
- g. Among workers aged 22–25, hiring into AI-exposed occupations has declined. A 14% decline in job entry rates for young workers in highly exposed occupations..
- This may indicate that companies are:
Automating entry-level tasks
Reducing hiring for junior roles
Using AI to increase productivity of existing employees

However, researchers caution that the trend is still statistically weak and requires further monitoring.

Possible Future Labour Market Scenarios

Economists outline several possible outcomes of AI diffusion.

1. Job Augmentation: AI enhances worker productivity while humans retain decision-making roles.

2. Job Transformation: Tasks change significantly but employment remains stable.

3. Job Displacement: Automation replaces workers performing routine cognitive tasks.

4. Creation of New Occupations: Technological innovation generates new industries and job categories.

The current evidence suggests that augmentation and transformation dominate in the short term.

E.g. The Internet Revolution created entirely new industries while transforming existing jobs rather than eliminating them. AI may follow a similar pattern.

Policy Implications

Workforce Reskilling: Governments must invest in digital skills and AI literacy to prepare workers for technological change.
Education Reform: Curricula should emphasise creative, analytical, and interpersonal skills less vulnerable to automation.
Monitoring Labour Markets: Continuous measurement frameworks such as observed exposure can help policymakers detect early disruption.
Supporting Young Workers: Entry-level job opportunities may decline in AI-intensive sectors, requiring targeted support.

Read More Notes
Environment Notes	Art and Culture Notes
Science and Tech	History Notes
Geography Notes	Indian Polity Notes
General Knowledge	International Relation