The key facts
- ECB Vice President Boris Vujčić noted that vulnerabilities have widened beyond overvalued stocks — to debt financing and operational risks
- Fed Governor Michael Barr warned of synchronized AI trading strategies and the possibility of faster bank runs
- Both regulators nevertheless acknowledged AI's potential to improve the efficiency of financial services
Warnings at two conferences
The European Central Bank and the US Federal Reserve on Monday issued synchronized warnings that the rapid spread of artificial intelligence could threaten financial stability — through cyberattacks, market manipulation, and increasingly synchronized trading, Anadolu reports. The remarks were delivered at Euro Finance Week in Frankfurt and at a conference on AI in finance in Milan.
ECB: risks have gone beyond overvalued stocks
ECB Vice President Boris Vujčić said that AI-related vulnerabilities have widened beyond stretched stock valuations and now include debt financing, concentrated positions, and operational risks. According to him, the likelihood of a market correction has increased, and banks' and financial intermediaries' dependence on a small number of technology providers creates additional vulnerabilities. “In particular, increasingly capable and autonomous AI systems amplify the scale, speed, and sophistication of cyberattacks,” the outlet quoted him as saying. Vujčić added that social engineering attacks are becoming more effective and may give attackers an advantage in the near term, while for financial institutions deploying defensive AI systems is more costly and comes with limitations. He also warned that a reassessment of AI expectations could trigger fund withdrawals from investment funds and forced asset sell-offs, and that leverage and concentrated portfolios could amplify market stress.
Fed: synchronized strategies and faster bank runs
Federal Reserve Board member Michael Barr, speaking in Milan, noted that traders using similar AI models trained on similar data may adopt synchronized strategies, increasing the risk of market disruptions. Shared weaknesses in AI systems, he said, can lead to failures with systemic consequences, and malicious actors could manipulate information while simultaneously deceiving multiple trading systems. Barr cited research showing that some self-learning algorithms may develop collusive behavior or learn to manipulate markets without explicit instructions. He noted that generative AI is not yet widely used for autonomous trading, but urged regulators and market participants to prepare for broader adoption. A separate warning concerned the possibility that AI agents could accelerate bank runs by automatically transferring customer funds in response to rate differences or signs of instability. At the same time, Barr acknowledged that AI can improve access to financial services and increase efficiency: “I believe we can and must ensure that this happens in ways that protect users and support financial stability,” Anadolu quoted him as saying.

