AI in action · Regulation
UK's Top Bank Supervisor: Lenders Are Trimming AI Risk
UK banks have started cutting their exposure to AI, the PRA's Katharine Braddick said, but the Bank of England has no plan to force it through capital rules.
"We're all alarmed at the moment, aren't we?" Katharine Braddick, chief executive of the Prudential Regulation Authority, the Bank of England arm that supervises UK banks and insurers, asked in her first interview since taking the job in July, published by Bloomberg on November 10, 2026. Her answer to her own question was measured: UK banks have begun reducing their AI exposure amid concern about a sharp correction in the sector.
She described it as ordinary credit discipline rather than a retreat. Banks are reassessing their risk appetite and deciding which borrowers to keep — "Yes this client, not that client," in her words. The Bank's review of lenders' AI exposure has not found that they would amplify an AI shock, and it is not planning to use capital requirements to push them toward smaller positions.
That puts the UK in a different place from the debate in the United States, where the Federal Reserve's November report counted more than $3 trillion of off-balance-sheet ties to AI computing. Braddick's point is that exposure is being managed one client at a time, not ignored, and that "operational resilience doesn't deliver financial resilience": keeping systems running is not the same as being able to absorb losses.
More is coming. The PRA and the Financial Conduct Authority said they will publish guidance later in November on risks from frontier AI and quantum computing, and the Financial Policy Committee's capital decisions are due on November 26. For bank shareholders, the interview is a reminder that the AI cycle reaches them through loan books as well as through the tech stocks in their portfolios.
Sources: Insurance Journal (Bloomberg) — Top UK regulator says banks are trimming AI risk as alarm grows (Nov 10, 2026)