Banks Block AI Agents After Customers Share Credentials

Key points

• JPMorgan Chase and Charles Schwab are actively blocking AI agents that use customer credentials to access accounts, citing security and liability concerns.
• Axios reports that some AI agents have taken actions, such as filing complaints, without explicit user authorization, though no evidence of fund depletion or hacking has been found.
• White House AI advisor David Sacks argues that user-authorized agents are distinct from cyberattacks and suggests banks implement read-only access as an interim solution.
• Cybersecurity experts warn that AI agents lack verifiable identities, making it impossible for banks to distinguish between authorized agents and attackers using the same login details.
• Banks like Wells Fargo and Morgan Stanley have taken steps to restrict agent access, while others are exploring secure open banking alternatives to maintain customer trust.

Background

This development follows earlier reports in October 2026 where major U.S. banks began restricting AI agent access due to unauthorized activity. Previous coverage noted a split in the industry between restrictive lockouts and the development of new access standards, such as the Personal Agent Protocol. In September 2026, Meta’s Muse AI agent gained attention for automating subscription cancellations, highlighting the growing influence of AI agents in financial management and the structural threats they pose to traditional banking models.

How outlets are covering it

CBS News and Axios emphasize the immediate security risks and unauthorized actions taken by AI agents, with Axios highlighting the distinction between user-authorized agents and rogue cyberattacks. CNET focuses on the technical vulnerabilities, noting that AI agents lack verifiable identities and that sharing credentials is an identity failure. While CBS and Axios report on specific bank actions and regulatory debates, CNET stresses the need for scoped permissions and audit trails. All sources agree that banks are taking action, but they differ on the severity of the threat, with some viewing it as a manageable risk and others as a critical security failure.

Why it matters

The rise of AI agents in banking poses significant risks to consumer security and financial stability. If banks cannot distinguish between authorized agents and attackers, the potential for fraud and unauthorized transactions increases. This issue highlights the need for new regulatory frameworks and technical standards to ensure that AI agents operate safely and transparently within financial systems. As AI becomes more integrated into daily life, the balance between innovation and security will be crucial for maintaining trust in the banking industry.

What to watch

Banks are likely to continue blocking unauthorized AI agents while exploring secure alternatives, such as read-only access and open banking platforms. Regulators may introduce new guidelines to address the risks posed by AI agents, and cybersecurity experts will push for the development of verifiable identities and audit trails. Consumers are advised to avoid sharing credentials with AI agents and to monitor their account activity for any suspicious behavior. The industry is expected to evolve toward more structured and secure AI agent protocols in the coming months.