Federal Reserve Governor Michael S. Barr used a November 12, 2026, address to examine how artificial intelligence could change American inequality and work. His speech did not provide a forecast of jobs lost to automation. Instead, it highlighted a more difficult question for employers and policymakers: whether new capabilities will mainly help existing workers perform better or allow firms to substitute technology for tasks that previously required people.

The Federal Reserve has published Barr's remarks, titled American Inequality and the Future of AI. The governor's emphasis is on uncertainty, the timing of adoption and how widely any economic benefits are shared. These are different questions from whether an AI model can complete a particular benchmark or automate one demonstration task.

Assistance is not the same as displacement

Barr noted that actual worker displacement from AI has so far been relatively limited and that research also finds examples of improved productivity, stronger demand and new kinds of work. Some tasks may be easier to automate when they have clear rules and predictable outcomes, while work relying on human judgment, coordination, relationships or creativity may change more slowly or use AI as an assistant.

It is important not to turn task-level exposure into a headline about whole occupations disappearing. A job can contain dozens of activities, only some of which can be automated reliably. Organizations may choose to redeploy employees, grow output, change hiring plans or eliminate certain roles. A credible labour-market assessment must track those outcomes separately rather than assume one universal path.

The inequality question

Barr also discussed how existing differences in income and wealth shape the stakes. A worker depends primarily on wages, while ownership of businesses and other capital can generate income from profits and asset appreciation. When a technology raises productivity, the benefits need not be distributed evenly between those groups. Competition between technology providers, the availability of tools to smaller employers, and the choices made by firms all influence who captures the gains.

For employees, the pace of change matters. A gradual shift can give people time to learn new tools or move into different roles. A rapid transition may leave too little time to adjust, even if the technology eventually creates productive new activities. This is why preparation can be sensible without asserting that mass unemployment is already underway.

What should be measured

Future reporting should distinguish measured job losses from hiring freezes, changed tasks, higher productivity and new occupations. Researchers also need to look at which workers benefit, how wages evolve, and whether small businesses can access useful AI at costs they can bear. The social impact will be determined by deployment decisions and institutions as well as model capability.

The speech represents Barr's views, not a new decision by the Federal Open Market Committee. Its policy message is a call to evaluate and prepare for possible disruption while the evidence remains incomplete, rather than to treat any one forecast as inevitable.

Primary source: Federal Reserve, American Inequality and the Future of AI, November 12, 2026.