Citi, Bank of America and retailers block AI agent account access

By Splitfeed EditorialLast updated Oct 18, 2026, 17:20 UTC

Starting around mid-October 2026, Citi and Bank of America moved to restrict AI agents from accessing customer accounts, a step that puts two of the largest US banks alongside major retail platforms that had already taken similar action. Amazon and eBay are among the retailers named as having blocked third-party AI tools from interacting with their platforms. The restrictions target a specific category of software, agents designed to act autonomously on a user's behalf, placing orders, moving funds, or managing accounts without direct human input at each step.

The significance here is less about any single institution and more about what the simultaneous moves suggest. AI agents capable of acting on financial and shopping accounts have been pitched as a next practical frontier for consumer AI, but the platforms and institutions that hold the underlying accounts have now begun drawing a clear boundary. Whether these decisions rest on liability concerns, fraud prevention, terms-of-service enforcement, or some combination is not yet spelled out in available reporting, and that ambiguity is itself part of what makes the dispute meaningful. The companies controlling the infrastructure are, in effect, deciding which AI capabilities can be built on top of them.

No published reactions from the Pro-AI, Anti-AI, or Middle Ground camps have surfaced yet for this story. Typically, Pro-AI voices would argue that these restrictions protect incumbent platforms from competition and slow a form of automation that genuinely benefits consumers. Anti-AI commentators would likely frame the blocks as a reasonable and overdue safeguard, pointing to the risks of autonomous agents making irreversible financial or purchasing decisions without clear accountability. A Middle Ground position would probably acknowledge both the security rationale and the concern that blanket bans could foreclose legitimate uses, calling instead for standardized access frameworks with clear liability rules.

The questions most worth watching are whether the banks and platforms publish formal policy language explaining the technical or legal grounds for these restrictions, and whether any AI agent developers respond with legal challenges or negotiate access agreements. A regulatory statement from financial oversight bodies would also shift the terms of the debate considerably.