The Bank of England is launching a review of artificial intelligence and machine learning amid concerns that its use could pose financial stability risks.
AI has been used in financial services for at least a decade, but adoption has recently accelerated thanks to breakthroughs in the use of Large Language Models.
In its half-yearly Financial Stability Review, the BofE says AI and ML could deliver significant benefits to the UK financial services sector, by driving greater operational efficiency, improving risk management, and providing new products and services.
However, wider adoption of the technologies "could conceivably also pose system-wide financial stability risks, for example by amplifying herding or broader procyclical behaviours or increasing cyber-risk and interconnectedness," says the report.
Therefore, the bank says it will consider the financial stability risks of AI and ML in 2024 and seek to ensure that the UK financial system is resilient to these risks.
Andrew Bailey, BofE governor, told a press briefing: "We obviously have to go into AI with our eyes open."
"It’s not out of control in the sense of 2001: A Space Odyssey. It’s actually that the thing is so complicated in many of its forms that understanding exactly what the black box delivers is very hard."
To illustrate the point he highlights ' “hallucinations” – unexpected creations from generative AI. “You can’t have that sort of thing happening” in financial firms, Bailey says.