BoE's Bailey tells G30 they 'have got to address' AI agent concerns
- BoE's Bailey warns 'agentic' AI could mislead consumers into thinking their digital minions are legal persons, creating a liability swamp when trades go sideways.
- Global regulators, from the IMF to the G7, are collectively sweating a future where a clever black box drains your bank account and the only response is a shrug.
- Wall Street is already slamming the brakes: big banks are banning AI agents from touching client accounts over fears they'll chase higher yields and set off a deposit stampede.
- Bailey suggests 'model testing' over regulation, because nothing says financial stability like a stress test on a hallucinating algorithm.
- The Bank of Korea's governor offers a chilling comfort: the old rules might apply, but AI is a 'very, very clever black box'—so good luck with that.
Brief Summary
Bank of England Governor Andrew Bailey has officially declared the age of the 'AI agent' a financial stability headache, warning policymakers at a Group of Thirty shindig in Bangkok that folks are dangerously deluded into thinking their automated financial gofers are legally accountable. He's got a point: you can program an 'economic agent' to move your cash, but you can't make it a 'legal agent'—so when it does something idiotic, the blame game becomes a bureaucratic nightmare.
Why This Matters
You may think you're just automating your savings, but Bailey is highlighting a legal minefield: if your AI agent goes rogue and drains your account or triggers a market panic, you're left holding the bag while regulators argue over definitions. The bigger threat is systemic—if left unchecked, these black boxes could amplify a financial crisis the same way subprime mortgages did, only faster and with less transparency. The G7 is now proposing 'tabletop exercises' to simulate AI-driven cyber attacks, which is comforting right up until you realize that's an admission they're flying blind.