What we know

American Express has introduced a new framework designed to prevent AI agents from making unauthorized purchases on behalf of users. This system aims to ensure that only transactions explicitly approved by the user are processed, reducing the risk of errors or disputes as AI-driven payments become more widespread. The framework specifically targets unauthorized purchases initiated by AI agents. It requires explicit user authorization before processing any AI-initiated transaction.

Why it matters

As AI-powered personal assistants and agents increasingly handle financial transactions, the potential for unauthorized or mistaken purchases grows. American Express’s framework addresses these emerging risks by adding verification steps to confirm user consent before completing payments initiated by AI. This development reflects a broader push within the financial industry to adapt payment security measures to the challenges posed by AI-driven commerce.

What is still unknown

Details about the framework’s accuracy, which languages it supports, and when it will be widely available have not been disclosed. No independent tests of its effectiveness or security have been reported.