Artificial intelligence is moving beyond recommending products. A growing group of services is being designed to compare choices, select an item and complete a purchase for the user. The idea could save time, but it also asks consumers to hand an automated system access to personal preferences, account information and money.

Six banks, including Bank of America and Capital One, published joint principles Sept. 22 for what they call trusted agentic commerce. The term describes transactions in which an AI agent acts for a consumer or business during shopping and payment. The other participating institutions are ASB Bank, Commonwealth Bank of Australia, ING Group and NatWest Group.

The banks said the technology could reshape how people choose and pay for products and services. They also identified unresolved questions involving identity, authorization, fraud prevention, liability and consumer protection. Their principles call for consumers to remain in control, understand what an agent is doing and have confidence that their data and payments are secure.

That is different from using a chatbot to brainstorm gift ideas or compare features. Research assistance leaves the final decision with the shopper. An agent that can place an order must know which product to choose, how much it may spend, which seller is acceptable, where the purchase should be shipped and which payment method it may use.

A shopper could ask an agent to find an appliance within a budget, compare delivery dates and order the best match. But the agent could misunderstand the request, select a sponsored recommendation, use an untrusted seller or send card information through a process with weaker protections….READ MORE….