Wall Street Research Warns AI Agents Could Spark Bank Runs
A major Wall Street firm warned that autonomous AI agents could cause a new type of bank run by moving deposits and trading flows rapidly. The research modeled simultaneous reactions to market signals, potentially straining settlement and forcing banks to seek emergency liquidity or cut lending. It recommended real-time monitoring, transfer limits, stronger authentication, and stress tests for correlated automated withdrawals.
The warning came in a research note from a large Wall Street institution. It focused on software agents that handle transactions and money management for users. Simulations assumed many popular agents would react at once to the same signal, such as falling commercial paper yields or a credit event. Funds could move from deposits and money market balances into assets or platforms viewed as safer within minutes or hours.
Institutional agents could also speed withdrawals from money market funds, prime cash funds, and short-term funding conduits tied to corporate financing. The note urged real-time monitoring, transfer throttles, stronger machine authentication, and stress tests covering correlated automated withdrawals. It also raised accountability questions among banks, fintech platforms, and AI providers.
Depositors, investors, banks, fintechs, and firms relying on short-term credit may be affected if automated withdrawals accelerate during stress. Households could face delayed access to cash or tighter lending if banks conserve liquidity. Institutions might adopt transfer caps or stricter authentication, which may add friction for legitimate users. Regulators and AI providers could face pressure to clarify responsibility and set safeguards. The broader effect may depend on whether guardrails keep speed and correlation manageable.