Back to Home

AI in the labor market: risks and data 2026

The study assesses the susceptibility of 800 professions to AI based on O*NET, Anthropic and Eloundou et al. A slowdown in hiring young people in vulnerable areas has been identified without unemployment growth. The gap between potential and implementation remains significant.

AI Risks for Professions: Fresh Data from the USA
Advertisement 728x90

AI and the Job Market: Quantifying Occupational Vulnerability

Research reveals a gap between AI's theoretical potential and its actual use across professions. Using O*NET data, the Anthropic Economic Index, and Eloundou et al. (2023), we've assessed vulnerability for 800 U.S. occupations. Digital roles are most at risk, where AI is already in partial use but covers less than 20% of possible tasks.

High-vulnerability jobs are projected to grow more slowly through 2034, per BLS forecasts. They tend to employ more women, workers over 45, those with college degrees, and above-median earners. Unemployment hasn't spiked systematically since 2022, but hiring of young workers (22–25) has slowed by 14%.

Methodology for Calculating Vulnerability

Vulnerability = observed + theoretical components.

Google AdInline article slot
  • *ONET**: Task descriptions for 800 occupations.
  • Anthropic Economic Index: Current AI adoption levels.
  • Eloundou et al. (2023): Potential for task time reduction.

Observed vulnerability: Full automation of tasks = weight 1, assistive use = weight 0.5. Theoretical vulnerability estimates acceleration potential without implementation barriers. About 30% of workers are in zero-vulnerability roles: manual labor like cooks, mechanics, and firefighters.

Top Vulnerable Occupations

Highest-risk leaders:

  • Programmers — heavy AI coverage in coding and debugging.
  • Customer service reps — chatbots and automated query handling.
  • Data entry operators — processing and verification.

In these fields, AI speeds up routine work but doesn't fully replace humans. CPS (Current Population Survey) charts show steady unemployment since 2016. The pandemic drove layoffs in low-vulnerability service jobs.

Google AdInline article slot

Employment Trends After ChatGPT

Since ChatGPT's November 2022 launch, unemployment in digital roles has held steady. The big shift is in hiring:

  • Ages 22–25: Down 0.5 percentage points (14% from 2022 baseline).
  • Over 25: No change.

This signals caution: Companies favor experienced hires over entry-level. AI's theoretical potential (up to 50% of tasks) rolls out slowly due to adoption hurdles.

| Group | Hiring Change (22–25) | Unemployment |

Google AdInline article slot

|--------|-----------------------|--------------|

| High Vulnerability | -0.5 p.p. (-14%) | Stable |

| Low Vulnerability | No change | Stable |

Key Takeaways

  • Current AI use is a fraction of what's technically feasible.
  • Slower job growth in vulnerable fields, per BLS.
  • Risk demographics: Women, older workers, educated professionals.
  • No mass layoffs, but higher barriers for young entrants.
  • Wide gap between AI theory and real-world rollout.

Monitoring Outlook

CPS data updates monthly for trend tracking. Future analyses will factor in AI agents and new metrics. For developers and managers: Prioritize low-automatability skills — creativity, system integration, AI ethics. The job market adapts via reskilling, not mass firings.

— Editorial Team

Advertisement 728x90

Read Next