Six Global Banks Warn of Risks in AI-Powered Agentic Shopping

On September 22, 2026, six major global banks, including Bank of America and Capital One, jointly released a report titled "Building Trust in Agentic Commerce," warning that agentic shopping carries risks spanning transparency, safety, privacy and data, choice, and interoperability, which increase as greater autonomy is given to AI agents.

The report says consumers are unclear whether AI agents will act in their interests and are concerned that agents may buy the wrong items, overspend, or lose money to scams and fraud. It also notes that AI agents could prioritize certain products or payment methods due to incentives such as higher commissions and lower token costs, rather than choosing what is best for the customer.

Agentic commerce also creates potential for higher rates of scams, fraud and disputes, and any data breach could jeopardize sensitive data belonging to consumers and merchants relying on these agents. The report cites Meta's recent fix of a zero-day vulnerability in its AI assistant Muse, which could allow an attacker to hijack the assistant and exploit user-granted permissions, as an example of the underlying risks.

The banks are not dismissing agentic commerce outright, saying it could become a mainstream way for consumers and merchants to transact and expressing eagerness to work with customers, industry and stakeholders on its future. However, they stress that trust must first be built through industry standards, policies and consumer protections, which currently lag technological developments. The consortium is now working on a follow-up paper detailing how the five key principles outlined in the report can be implemented.