Federal Reserve Governor Michael S. Barr warned Nov. 16 that AI agents could accelerate deposit runs, while European Central Bank Vice President Boris Vujčić cited market and cyber risks with implications for financial stability.
Barr said traders using a small group of AI models trained on overlapping data could respond to new information in similar ways. A flaw shared across systems or manipulated inputs could then disrupt markets more broadly. Generative AI was not yet widely used for autonomous trading, he said.
Barr also cautioned that agents might detect yield differences or signs of financial strain and move deposits automatically, leaving banks and regulators less time to respond. Major banks, he said, had found agents trying to act in retail accounts without verifiable authorization or confirmation that the actions matched customers’ instructions.
Vujčić said AI-related funding is spreading from equity markets into debt markets. If market expectations about AI shift, he said, funds’ AI holdings could lose value. Redemptions and forced sales could deepen pressure when portfolios are concentrated or leveraged.
He also pointed to financial institutions’ reliance on a limited group of technology vendors for critical services. More capable AI could give cyber attackers greater speed and capability, he said, while banks may face higher costs and constraints as they use AI for cyber defense.
Barr said AI could improve access to financial services and efficiency. He compared debt funding for the AI buildout with the telecom investment cycle of the late 1990s and early 2000s; Vujčić noted that past IPO surges have often coincided with market peaks.
The ECB’s latest Financial Stability Review was due Wednesday, Nov. 18, Vujčić said.


