AI, Work and the Economy

Latest/Story

The Federal Reserve on AI

Six items, from 28 Sep 2026 to 16 Nov 2026, oldest first.

See also: Point 3 of our view: Pay's share of income falls · Point 2 of our view: Fewer office jobs, slowly at first

Economy and marketsPolicy

Fed's Cook (28 Sep): AI may already cut demand for coders, translators, entry-level staff; rate cuts may not fix it

Fed Governor Lisa Cook said AI 'could bring the most significant reorganization of work in generations', that it may be reducing labour demand in software coding, translation and entry-level tasks, and that cutting rates against AI-driven, skills-mismatch unemployment 'could risk fueling inflation', with inflation at 3.8% and power and water costs up about 5%.

Why it matters. A Fed governor says AI may already be cutting some jobs and that lowering interest rates might not help those workers without fuelling inflation.

Next. US job openings on 29 Sep and payrolls on 2 Oct.

Sources: federalreserve.gov, news.google.com

Economy and marketsPolicyUpdate

Fed's Barr (29 Sep): 'further policy adjustments are likely' after September hike; AI may substitute for young workers

Governor Barr said inflation has been above 2% for five and a half years, the labour market is in rough balance at 4.1% unemployment, and more tightening is likely needed; he added that AI already looks like a labour substitute for younger, less experienced workers and that a lasting productivity boost would mean higher equilibrium rates.

Why it matters. A Federal Reserve governor says more rate rises are likely and that AI may already be replacing young workers, while also arguing that an AI productivity boom would keep interest rates higher for good.

Next. The Fed's next decision on 28 October; US September jobs report on 2 October.

Sources: federalreserve.gov

PolicyUpdate

Fed minutes (7 Oct): most officials see another hike 'likely' by year end; AI build-out 'surprised to the upside'

Minutes of the 15-16 September meeting, which raised the federal funds range to 3.75-4% by 12 votes to 0, say most participants judged another increase likely appropriate by year end. Several said the AI build-out's scale and pace kept surprising to the upside, and some warned AI-related demand could push demand above supply; staff put August PCE inflation at 3.8%, core 3.4%.

Why it matters. Most US central bankers expected to raise rates again before year end, and several described investment in AI as a source of demand large enough to keep inflation up. If they act on this while hiring stays weak, higher borrowing costs could meet a labour market with little slack to absorb them.

Next. US inflation for September in mid-October and the Fed's decision on 28 October.

Sources: federalreserve.gov

Economy and marketsReportUpdate

Fed Governor Barr (12 Nov): AI may be 'capital-biased'; labour's share hit a record low 52.8% in Q2 2026

In a speech at Yale on 12 Nov, Fed Governor Michael Barr said AI could be innovation whose returns go mainly to capital, and that if a few firms capture them labour's share would fall and inequality rise. He cited labour's share at a record low of 52.8% in Q2 2026, falling since 2001, and early signs of AI substituting for younger workers.

Why it matters. A member of the Federal Reserve's Board said AI could send its returns mainly to the owners of capital rather than to workers, citing labour's share of income at a record low of 52.8% in the second quarter. If central bankers come to read the fall this way, policy debates on taxes, transfers and rates might begin to shift before the data settle the question.

Next. The revised third-quarter productivity and labour cost figures on 4 December.

Sources: federalreserve.gov · Point 3 of our view · Context: not counted as evidence

AI and technologyNew dataUpdate

Fed Vice Chair Jefferson (13 Nov): about 25% of US firms use AI; its productivity contribution so far 'limited'

Jefferson cited Census data showing about a quarter of US firms using AI in at least one business function (about 40% in knowledge-intensive sectors), productivity growth of 2.2% a year in 2020-25 against 1.5% the cycle before, and a survey putting AI's lift to sales per employee at about 0.5% over three years. He called the labour market 'stabilized' at 4.1% unemployment and named the risk that AI 'could displace workers much faster than they can adapt', some of it structural.

Why it matters. A senior Federal Reserve official puts the share of US firms using AI at about one in four and finds its effect on productivity so far small, while warning that it might displace workers faster than they can retrain. If adoption keeps spreading at this rate, job losses could remain modest for some time; a sudden rise in adoption would be the condition for them to come faster.

Next. The US jobs report for November, due 4 December.

Sources: federalreserve.gov · Point 2 of our view

PolicyReportUpdate

Fed Chair Warsh (16 Nov) sets out three AI scenarios, including job losses above full employment and a bust

Opening the Fed's productivity and jobs task force conference in San Francisco on 16 Nov, Chair Kevin Warsh named three illustrative scenarios: broadly shared gains; automation displacing jobs faster than it complements them, with growth above trend but unemployment above full employment; and disappointing returns on AI investment, falling equity valuations and a slowdown that raises unemployment. He said Board staff are building model-based scenarios and that productivity has not yet clearly moved to a higher trend.

Why it matters. The head of the US central bank said publicly that AI could push unemployment above full employment even while the economy grows, or that disappointing returns on AI investment could bring a slowdown. That the Fed is preparing for both may shape how quickly it responds if either begins.

Next. The Fed's December rate decision and its minutes on 25 November.

Sources: federalreserve.gov · Point 7 of our view · Context: not counted as evidence