US EditionAnthropic Is Tracking Which Jobs Are Most Exposed to AI. These 10 Professions Top the List.
Anthropic's new metric shows about 75% AI task exposure in programming jobs and reveals AI use is much lower than its theoretical potential, with no major job losses yet.
- Updated on March 6, 2026, Anthropic, maker of the Claude chatbot, launched an early-warning system tracking which U.S. jobs face AI exposure, finding many white-collar roles near the front lines.
- Using three data sources, the study combines O*NET occupational task descriptions, task-level large language model capability estimates and real-world AI usage datasets to inform researchers and policymakers.
- Ranking occupations, Anthropic found computer programmers lead at 75% exposed, with customer service representatives at 70%, data-entry keyers at 67%, and market research analysts, sales representatives, and financial analysts also highly exposed.
- Researchers find that hiring among younger workers into exposed occupations has fallen by roughly half a percentage point, with a job-finding rate decline of about 14% versus 2022, yet labour-market unemployment indicators do not yet show broad disruption.
- Using BLS forecasts for 2024–2034, the analysis shows the highly exposed worker group earns about 47% more and includes more workers with graduate degrees, though occupations with higher exposure tend to have slightly weaker projected growth.
15 Articles
15 Articles
Programming, customer service, data entry roles among jobs most at risk from AI: Anthropic
A recent Anthropic study reveals that nearly 70% of tasks for programmers, customer service representatives, and data entry operators are already automated, placing these roles at high risk of displacement by large language models. The research highlights the growing real-world impact of AI on professions, with observed exposure metrics tracking actual AI integration in the workplace.
Anthropic publishes a new method for assessing the impact of artificial intelligence on employment, based not only on theoretical data on AI performance, but also on data on how technology is actually used.
Anthropic has developed a new measure that combines theoretical AI skills with real usage data. The result: programmers and customer service employees are most exposed, but unemployment in affected professions has not so far increased. Only young workers have first warning signals. The article "AI steals jobs": Anthropic study finds no evidence of frequent claims on the labour market first appeared on The Decoder.
Coverage Details
Bias Distribution
- 50% of the sources are Center, 50% of the sources lean Right
Factuality
To view factuality data please Upgrade to Premium
