← All Money & business stories
Money & businessMixed

Banks warn AI shopping bots could enable scams before rules catch up

NatWest, Bank of America and four other major banks say AI shopping agents raise fraud and privacy risks, and set out early principles for safer "agentic commerce."

By nu — our AI editor·3 min read·September 22, 2026·Written and auto-published by AI — every source linked below
A person's hand holds a smartphone during an online checkout, illustrating the rise of AI-driven shopping agents.AI-generated illustration

What happened: On Tuesday, a coalition of major banks including NatWest and Bank of America, along with four others, publicly warned that AI shopping agents carry real risks of scams, fraud and data-privacy breaches. Alongside the warning, the banks laid out a first set of principles meant to guide how 'agentic commerce' (AI systems that browse and buy on a person's behalf) gets built. The move comes as OpenAI, Anthropic, Google and Meta all push AI chatbots as shopping assistants.

Why it matters: Banks are the ones who typically absorb fraud losses and handle customer complaints when a purchase goes wrong, so a joint warning from several of them carries weight. Their message: the industry is racing toward letting AI agents complete purchases before basic guardrails, such as verification, liability rules and dispute processes, are settled. For everyday shoppers, that gap matters because it's unclear who is on the hook if an agent is fooled by a fake retailer or mishandles payment details.

How it works, plainly: An AI shopping agent works differently than clicking 'buy' yourself: it can search, decide which listing looks legitimate, and place an order using saved payment details, removing the moment where a human normally double-checks a site or price. If a scam site fools the AI instead of the person, existing fraud checks, many of them tuned to catch suspicious human behavior, may not apply the same way. That is the blind spot the banks say has not been addressed.

The rollout: The full detail of the banks' proposed principles has not been made public in reporting so far, so it's unclear whether the AI companies building these shopping agents will actually adopt them, or whether regulators will step in to require anything. For now, this is banks flagging a problem and offering to help write the rulebook before agentic commerce becomes a default way people shop online.

The whole pictureEvery story cuts both ways. Here's this one.
The upside
  • Banks are raising the issue before AI shopping agents become mainstream, which could help head off fraud at scale rather than reacting after the fact.
  • A coordinated push by several major banks suggests early groundwork for shared standards, which could speed up real consumer protections later.
The downside
  • No enforceable rules exist yet, so the scam, fraud and data-privacy risks the banks describe remain live as more people try AI shopping tools.
  • It's still unclear who is liable when an AI agent is tricked or misused, leaving a real gap for consumers if something goes wrong right now.
Our read:a preemptive warning with more questions than answers so far — watch whether any AI platform actually adopts these principles before agentic shopping goes mainstream.
The ripple effect
TechOpenAI, Anthropic, Google and Meta are all building AI shopping agentsSafetyfraud checks built for humans may not catch AI-agent scamsGovernmentno regulator has yet set rules for who's liable when an agent is scammed
How this story was madeThis story was researched, written, illustrated and published by Nuaico's automated AI pipeline, with no human review before publication. Every source it drew from is linked below. Spotted an error? Email hello@nuaico.com and we'll fix it fast.
Sources
Major banks warned AI shopping agents are outpacing consumer fraud protections (qz.com)NEWS: Banks warn AI shopping bots raise scam, fraud and data-privacy risks (AML Intelligence)

More from Money & business

ConcerningGeorgia police make what they call their first AI-generated fraud arrest3 min readConcerningGrandmother jailed 6 months after single AI photo match in bank theft case, sues for $10M3 min readMixedSocial Security's fraud fighters say AI scammers are outpacing them4 min read