Federal Reserve Warns AI Agents Could Trigger Bank Runs Within Minutes

Autonomous financial assistants expose new risks for banking security and stability.

MILAN, Italy — November 2026. Major American banks are detecting artificial intelligence assistants attempting to perform operations in customer accounts without institutions being able to reliably verify their identity, authorization or compliance with account holders’ instructions. The warning came from Federal Reserve Governor Michael S. Barr on November 16, raising concerns about a new generation of financial risks as autonomous AI systems become increasingly capable of conducting transactions.

Speaking at the Seventh Annual International Conference on Artificial Intelligence in Finance, Barr explained that AI agents differ from conventional chatbots because they can independently execute tasks, including purchasing products, analyzing financial information and transferring funds. Although customers may authorize some activities, banks cannot always determine whether an automated instruction reflects genuine consent or exceeds the authority originally granted.

The concern is already influencing banking practices. Institutions including JPMorgan Chase, Morgan Stanley and Citigroup have introduced restrictions on external AI agents accessing customer accounts. The objective is to prevent unverified systems from conducting financial operations while preserving legitimate opportunities for customers to use approved digital services.

Barr emphasized the importance of strong authentication, explicit authorization for specific transactions and technical limitations defining what an AI assistant is permitted to do. His remarks did not identify the affected banks or disclose how many suspicious incidents had occurred, and they did not establish that the detected attempts resulted in unauthorized transfers.

Beyond individual account security, the Federal Reserve official identified a potentially more serious systemic risk: autonomous agents could accelerate bank runs. During periods of financial uncertainty, AI systems monitoring interest rates or signs of institutional instability might automatically transfer deposits to competing banks, producing simultaneous withdrawals faster than conventional liquidity-management procedures could respond.

The warning also extended to the financing of artificial intelligence infrastructure. According to the Federal Reserve’s latest Financial Stability Report, debt held on the balance sheets of the seven largest AI companies is expected to exceed $800 billion before the end of 2026, while certain long-term contractual commitments and other off-balance-sheet obligations exceed $3 trillion. Barr cautioned that complicated financing arrangements could magnify market instability if investors substantially reassess expectations for AI profitability.

Financial markets present additional challenges. If investment institutions increasingly rely on similar AI models, their trading strategies could become synchronized, amplifying price movements during periods of uncertainty. Barr also cited research suggesting that certain self-learning algorithms can develop manipulative or collusive behavior without explicit instructions, although he acknowledged that generative AI is not yet widely deployed for fully autonomous trading.

Despite these concerns, Barr maintained an optimistic outlook regarding artificial intelligence’s long-term economic potential. He argued that the technology could improve access to financial services, increase productivity and strengthen market efficiency, provided institutions establish effective safeguards and maintain meaningful human accountability.

The emerging challenge is therefore not simply preventing malicious software from entering banking systems. It is ensuring that increasingly autonomous technologies remain transparent, controllable and aligned with the financial decisions their users actually authorize.

Where Innovation Meets Responsibility.

Related posts

OpenAI Expands ChatGPT Advertising Across Latin America With Local Spanish-Language Support

Sam Altman Praises Jensen Huang but Declines to Endorse His Nvidia Comeback

Nissan, Kia, Avatr and Volvo Unveil Future Mobility Technologies in Colombia