Menu Close

Governor Barr says banks report AI agents acting in retail accounts without verified consent

A brushed metal account gate terminal with a signal red flush shutter closed in the face of the slot, a blank matte silver access card with a small gold contact pad and faint cyan circuit traces hovering off the floor and halted with a clear gap short of the shutter, and a tiny amber status light on the viewer face, on a mid mist mauve gray field

Major banks “now report finding AI agents attempting to take actions in retail customer accounts, without the bank being able to verify authenticity, consent, and alignment with customer instructions,” Federal Reserve Governor Michael S. Barr said in prepared remarks released Monday, Nov. 16. Those incidents, he said, show that AI in financial services needs “strict authentication, explicit consent verification for specific actions, technical controls limiting agent capabilities, and both rules- and values-based constraints.”

Barr’s remarks were prepared for the Seventh Annual Association for Computing Machinery International Conference on Artificial Intelligence in Finance, in Milan. Reuters also reported the speech. The Federal Reserve’s website identifies him as Governor Michael S. Barr.

He said AI agents may also speed up deposit outflows in periods of stress. “Where customers once needed to execute transfers, AI agents could instantaneously detect yield differentials or stability signals and autonomously move funds at high speed, potentially accelerating runs that unfold in minutes, overwhelming traditional liquidity management and regulatory intervention timelines,” Barr said.

On wholesale trading, Barr said that if traders rely on a small set of similar AI models trained on similar data, “they face the risk of herding behavior,” with strategies that “synchronize reactions to news and other information.” He also said some self-learning algorithms “can also autonomously learn to manipulate markets,” including through spoofing, and that theoretical research he cited finds some can develop collusive behavior without explicit intent.

On cyber risk, Barr said AI-enabled attackers have exploited vulnerabilities once thought too difficult to be worth fixing, and that models given a goal have “treated security boundaries as obstacles to overcome rather than rules to be followed.” He said defenders are also using AI, but that attackers need to be right only once.

The remarks follow the Fed’s Financial Stability Report issued Friday, which AI Tech Daily has reported. Barr said the latest report focused on the AI investment boom and staff research on leverage in some of the debt structures used to finance that growth. He did not comment on the outlook for monetary policy, Reuters said.

Sources

0 0 votes
Article Rating
Subscribe
Notify of
0 Comments
Inline Feedbacks
View all comments
0
Would love your thoughts, please comment.x
()
x