UK banks cut AI exposure as Bank of England flags sector risks

UK banks are reducing their exposure to artificial intelligence as concerns grow about a possible correction in the sector, according to Katharine Braddick, head of the Bank of England’s Prudential Regulation Authority (PRA). Braddick said banks are reassessing their risk appetite and making selective decisions about AI-related clients. The central bank is not currently planning to use capital rules to push banks to lower their AI exposure but is monitoring the situation closely.

The PRA, together with the Financial Conduct Authority, will soon publish guidance on managing risks from advanced AI and quantum computing. Braddick said this guidance will set out governance expectations and aims to foster closer engagement between the regulator and the industry. She also referenced ongoing work on operational resilience, including the "minimum viable bank" framework, which would require banks to maintain core services during cyberattacks.

Capital requirements remain a focus of public debate, with the government urging the Bank of England to implement reforms quickly to support competitiveness. Braddick said stability and predictability are priorities for the PRA as it develops its first five-year strategy, which will include more transparency around regulatory decisions.

The PRA has also received enquiries from firms considering relocating to the UK, but Braddick said the regulator does not actively seek to attract business. The outcome of current capital reform discussions is expected to be announced on 26 November.