Michael Barr, a governor of the Federal Reserve, told the central bank's Next-Gen Financial Inclusion conference on Tuesday 14 July that "as of right now, there has been little evidence of economy-wide job displacement from AI".1 He built the case on published work rather than Fed staff modelling: a Brynjolfsson, Li and Raymond study, and a Noy and Zhang experiment finding college-educated professionals finished assignments 40% faster and 18% better with AI, with the largest gains going to the workers who performed worst without it.
Read the opener. "As of right now" describes what today's aggregate data shows and closes nothing. Barr is the governor who in March called the US labour market "low hire, low fire", a phrase this beat has since read as The Fed's tacit nod to the argument that AI suppresses hiring rather than causing redundancies. Nothing in the 14 July text withdraws it.
The stratification he disclosed matters more than the headline. AI adoption runs at 43% among workers holding graduate degrees against 10% among those with a high-school education or less, and the top-earning fifth of US households took 52% of 2024 income against 3% for the bottom fifth. A technology adopted four times more heavily by the already-advantaged does not distribute its gains evenly, whatever it does to the total. Barr named education, competition and tax policy as the remedies, and every one of them belongs to Congress, not to the Federal Reserve. A central banker who lists the answers and disclaims all three is describing the limit of his own instruments.
