How Could AI Affect Jobs?

AI can now perform parts of many jobs: drafting documents, writing code, handling customer service, summarizing records, and reviewing images.

Economists generally distinguish between tasks and jobs. Most jobs are bundles of tasks, and AI usually automates some of them rather than all of them. That can mean fewer workers doing the same output, the same workers doing more, or new roles that did not exist before.

Studies so far find measurable productivity gains in specific settings, especially for less-experienced workers in writing, support, and programming tasks.

Economy-wide effects are much harder to measure, and credible estimates of displacement vary widely. Official labor statistics have not yet shown a clear economy-wide break attributable to AI, though effects are visible in some occupations and entry-level hiring.

Proposals fall into a few buckets: workforce training and retraining, rules on how employers use AI in hiring and monitoring, disclosure when automated systems make employment decisions, and broader questions about the safety net.

Several states already regulate automated decision-making in employment. At the federal level, most action so far has been research, workforce programs, and guidance rather than binding rules.