Banks warn AI shopping bots increase fraud and privacy risks

By Splitfeed EditorialLast updated Sep 26, 2026, 20:18 UTC

On September 22, 2026, major banks issued warnings that AI shopping agents, autonomous tools capable of independently making purchases on behalf of users, create elevated risks of scams, fraud, and data-privacy breaches. The warnings, which Reuters and other outlets reported simultaneously, centered on a specific concern: that agentic commerce is advancing faster than the consumer protections and regulations designed to govern it. Alex Jimenez noted on Bluesky that the banks published a report flagging how agentic commerce could increase scams, fraud, and payment disputes, while Neira Jones added that six global banks, including Bank of America and NatWest, have responded by issuing principles intended to make agentic commerce safer.

The core dispute is about liability and architecture. When an AI agent is deceived by a fraudulent merchant, or mishandles payment credentials, it is not yet settled who bears responsibility, the consumer, the bank, or the platform running the agent. Dr. James Howard identified the structural problem precisely, arguing that agentic payments need to bind product, merchant, amount, and permitted payment method directly to the user's original intent, so that the agent cannot be redirected or manipulated mid-transaction. This is a design challenge without a broadly adopted solution, and existing consumer fraud protections were built for humans making deliberate choices, not for software acting autonomously.

The Anti-AI camp treats the bank warnings as confirmation of a pattern they have argued for some time. Accounts including London Startup Briefly and Privacy Professionals Briefly framed the situation as one where AI tools are outpacing regulation and shifting risk onto consumers. Quartz summarized the concern as consumers facing scams and data breaches as AI commerce tools advance faster than industry safeguards, and at least one reaction, from the account Nona ABOLISH ICE, treated the development as simply another accumulation of foreseeable harms. The Middle Ground camp acknowledges the problems but points to the banks' own principles as evidence that the industry is at least attempting to build guardrails; Valentine Interactive stressed that a clear path from AI recommendation to a verified website and human contact remains an essential safeguard. The Pro-AI camp has not yet published reactions to this story, though that camp would typically argue that agentic tools can be made secure and that fraud risks are manageable engineering problems rather than reasons to limit the technology.

The next concrete development to watch is whether the principles issued by Bank of America, NatWest, and their co-signatories translate into specific technical standards or regulatory proposals, and whether consumer protection agencies in the US or UK move to extend existing fraud liability rules to cover AI-initiated transactions.