BTC/USD $68,420 +2.8%
ETH/USD $3,540 +1.4%
SOL/USD $142.80 -0.6%
BNB/USD $605.20 +0.9%
XRP/USD $0.62 -1.2%
DOGE/USD $0.18 +5.4%
BTC/USD $68,420 +2.8%
ETH/USD $3,540 +1.4%
SOL/USD $142.80 -0.6%
BNB/USD $605.20 +0.9%
XRP/USD $0.62 -1.2%
DOGE/USD $0.18 +5.4%
Markets

Fed Warns AI Agents Could Trigger Bank Runs in Minutes as Major Banks Report Unauthorized Activity

Key takeaways Major banks report AI agents attempting unverified account actions Fed warns autonomous transfers could accelerate bank runs Regulators seek stronger authentication and transact

AnonymousCryptoCompass newsroom
November 17, 2026
3 min read
NEWS
Hero article visual / chart / editorial image
CryptoCompass editorial visual for markets coverage.

Key takeaways

  • Major banks report AI agents attempting unverified account actions
  • Fed warns autonomous transfers could accelerate bank runs
  • Regulators seek stronger authentication and transaction-level safeguards

Federal Reserve Governor Michael Barr warned that artificial intelligence agents could accelerate bank runs to the point where deposit withdrawals unfold within minutes, potentially overwhelming traditional banking safeguards.

Speaking at an AI finance conference in Milan on November 16, Barr also revealed that major banks have encountered AI agents attempting to perform actions in customer accounts without institutions being able to verify whether those actions were genuinely authorized.

The warning comes as financial institutions increasingly experiment with autonomous AI banking agents capable of processing transactions, managing accounts and making financial decisions without continuous human supervision.

Major Banks Report AI Agents Attempting Unverified Transactions

According to Barr's official Federal Reserve speech, major financial institutions are already encountering AI agents that attempt to interact directly with retail customer accounts.

The central problem is authorization. Banks may struggle to determine whether software genuinely represents the account holder, has permission to perform a particular action or is following the customer's original instructions.

Barr did not identify the institutions involved or disclose how frequently these incidents occurred. He also did not establish that the attempted actions resulted in completed unauthorized transfers.

Nevertheless, the findings highlight emerging security concerns as AI-powered payments become more common.

Unlike conventional banking software, autonomous agents can interpret instructions, select actions and interact with multiple financial services.

That flexibility creates additional risks when an agent misunderstands its instructions, accesses an account without sufficient authorization or responds to manipulated information.

AI Agents Could Turn Bank Runs Into Minutes-Long Events

Barr raised a broader concern about the speed at which autonomous financial software could move money during periods of market stress.

Traditionally, depositors must recognize potential problems, decide whether to withdraw funds and manually initiate transfers. AI agents could potentially perform those steps automatically.

Software monitoring bank stability or interest-rate differences might react to warning signals and move deposits before human customers even recognize a problem.

If many agents respond simultaneously, withdrawals could accelerate faster than banks and regulators can manage liquidity pressures.

The concern extends beyond unauthorized transactions. Even properly authorized AI agents could destabilize financial institutions if their automated strategies trigger simultaneous withdrawals.

Existing agentic payment systems already demonstrate how software can execute transactions based on previously established permissions.

Fed Calls for Stronger AI Safeguards Across Banking

Barr identified strict authentication, verification of consent for specific actions and technical restrictions on agent capabilities as essential protections.

His comments follow growing scrutiny of bank cybersecurity vulnerabilities as artificial intelligence expands into financial services.

The Fed governor also warned that automated trading systems could amplify market instability if numerous AI models react similarly to economic developments.

Separately, the Federal Reserve's latest financial stability assessment highlighted risks associated with the rapid expansion of AI infrastructure financing.

Banks are also evaluating the operating costs of these systems, with recent industry analysis raising questions about whether automation will deliver the anticipated savings.