AI's workforce impact: Study reveals which jobs are most likely to be replaced

A new analysis of 200,000 U.S. interactions with Microsoft Copilot maps each chat to the work activities defined in the U.S. O*NET database and then links those activities back to specific occupations. The authors compute an “AI applicability score” that reflects how frequently Copilot is used for an occupation’s tasks, how often those tasks are completed, and how broadly each task is covered within the job description.

An occupation earns a high score when Copilot reliably helps users finish core tasks that make up a large share of that job. For example, tasks such as information gathering, drafting text, and explaining concepts receive high completion rates; therefore, occupations dominated by these activities tend to rise to the top.

Based strictly on the data found inside the study, the ten leading jobs are:

• Interpreters and Translators – 98 percent of tasks overlapped with Copilot use
• Historians
• Passenger Attendants
• Sales Representatives (Services)
• Writers and Authors
• Customer Service Representatives
• CNC Tool Programmers
• Telephone Operators
• Ticket Agents and Travel Clerks
• Broadcast Announcers and Radio DJs

Most high-scoring roles depend on producing, tailoring, or conveying information; they are commonly found in sales, office support, media, education, and similar knowledge-intensive fields. By contrast, occupations centered on physical assistance, manual handling, or machinery operation—such as nursing assistants, massage therapists, and roofers—rank at the bottom of the list.

Today’s generative AI excels at tasks that involve words, data retrieval, and structured guidance; it is far less capable where hands-on skills or direct human contact are indispensable. As technology matures, its reach may broaden, but current evidence suggests the greatest near-term impact will be on roles where information exchange is the primary currency.