Amex unveils framework to block AI agents from unauthorized purchases
agents
| Source: TechCrunch | Original article
American Express is rolling out a framework to ensure AI agents only make purchases authorized by users, aiming to curb costly errors and disputes as AI‑driven shopping expands.
American Express announced a new framework aimed at ensuring that artificial‑intelligence agents can only complete purchases that users have explicitly authorized. The move comes as AI‑driven shopping assistants and bots become more common in e‑commerce, raising concerns that automated systems could inadvertently charge cards for items users never intended to buy. By embedding verification checks into the transaction flow, Amex hopes to block “costly mistakes and disputes” before they reach the consumer or merchant.
The development matters because it tackles a nascent risk at the intersection of fintech and generative AI. As voice assistants, chat‑based shoppers and other autonomous agents gain the ability to browse catalogs, add items to carts and finalize payments, the potential for mis‑aligned intent grows. Unauthorized charges not only erode consumer trust but also increase operational burdens for banks and merchants handling chargebacks. Amex’s framework signals that major card issuers are beginning to embed safeguards directly into their payment ecosystems rather than relying solely on post‑transaction dispute mechanisms.
Industry observers will be watching how quickly the framework is rolled out across Amex’s product lines and whether it will be adopted by other issuers. Key questions include how the verification process will interact with third‑party AI platforms, what standards will emerge for “authorized‑purchase” signals, and whether regulators will reference such safeguards in future guidance on AI‑enabled financial services. The next few months should reveal whether the approach curbs inadvertent spending and sets a precedent for broader industry adoption.
Sources
Back to AIPULSEN