AI estimates job displacement risk in the United States: high-skilled roles under threat
Andrej Karpathy, co-founder of OpenAI and former architect of Tesla Autopilot, conducted a study evaluating 342 professions in the US based on their likelihood of being replaced by artificial intelligence. The average risk score reached 5.3 out of 10, with nearly 60 million workers employed in high-risk sectors.
The findings suggest that the greatest exposure is not among low-skilled workers but rather among highly skilled and well-paid professionals. Programmers, financial analysts, and office administrators scored 9 out of 10, while lawyers received a score of 8.
Karpathy introduced a simple rule: if a job is performed entirely on a computer, the likelihood of automation is significantly higher. In contrast, physically demanding roles remain the least vulnerable, with professions like construction workers and roofers scoring just 1 out of 10.
Developers of the Claude model also report a slowdown in hiring young professionals in fast-moving industries, indicating early signs of labor market shifts.
Experts say these insights challenge traditional assumptions about job security and highlight the need for adaptation, as even highly qualified professionals may face increasing risks in the evolving AI-driven economy.
