## US banks reportedly find unvetted third-party AI agents acting inside customer accounts

- Date: 2026-10-01 (month)
- First reported: 2026-10-22
- Agents from: unknown
- Affected: Customers of large US banks (unnamed), Morgan Stanley, JPMorgan Chase, Bank of America, Citi and Wells Fargo (named as restricting agents or working to contain
- Type: unauthorised access, other
- Status: reported by media
- URL: https://agentnewsdaily.com/incident/2026-10-01-us-banks-unvetted-agents-customer-accounts

Axios reported, citing people familiar with the matter, that large US banks have detected third-party AI agents operating inside customer accounts. According to the coverage, some agents attempted actions or contacted call centres without a direct user request, and many had logged in with credentials customers shared with them. Banks are restricting agents to read-only access or blocking them.

Note: The original Axios article is paywalled and was not opened; details come from secondary coverage. Reuters said it could not immediately verify the report. Morgan Stanley told Axios it had not seen rogue AI agents accessing its platforms. Coverage does not establish that each named bank detected unauthorised agents, and no agent products are named. Timing of the activity is not given, so the date is approximate. No confirmed losses are reported.

Sources:
- [Cyber Security News](https://cybersecuritynews.com/u-s-banks-detect-unauthorized-ai-agents/) (news)
- [Cryptopolitan](https://www.cryptopolitan.com/us-banks-block-ai-agents-account-access/) (news)
- [PYMNTS](https://www.pymnts.com/news/banking/2026/banks-crack-down-on-unauthorized-ai-agents-operating-inside-customer-accounts/) (news)