Starting Tuesday, Wells Fargo will apply updated terms to millions of customers, including a warning aimed at people who use AI agents to automate online tasks. The bank’s message to early adopters is direct: “Use AI agents at your own risk,” according to its new terms.
AI agents that browse websites, complete online tasks and click on users’ behalf continued to stumble throughout 2026. The technology is still in its early stages, but a growing number of power users have begun relying on personal bots. While some early adopters have benefited from the efficiency of automation, others have faced public embarrassment when their agents made serious mistakes, including posting bank account details where everyone at work could see them.
Developers often market AI agents as tools that can make shopping decisions, from assembling grocery orders and finding vacation accommodations to locating a specific holiday gift for a parent. One user employed an agent to help purchase a new sofa, compare options within a set budget and track price drops.
Some users value these agents so highly that they may overlook the security risks of giving them account credentials and authorizing broad actions without carefully considering the consequences. Although autonomous tools still account for only a small share of online transactions, US banks are already pushing back against third-party AI agents.
Companies push back against autonomous agents
Banks are not alone. AI agents are also facing restrictions across the wider internet. Amazon, one of the largest online retailers, has blocked Meta’s Muse, one of the most widely used AI agents, from making purchases on its platform. Travel companies have objected as well: United Airlines recently said it would block individual bots from purchasing tickets on its behalf.
As the year draws to a close, several banks have introduced new policies or issued warnings about agents. In late October, JPMorgan Chase changed its terms to prevent third-party agents from logging in with customers’ credentials. On the same day, Bank of America warned customers not to share login information with agents. Alongside Wells Fargo’s announcement, Citibank and PNC also changed their terms and conditions to include warnings about agents.
Users approve goals, not every action
Automation remains a central selling point for always-on agents, but it is also where users can encounter problems. “Consumers approve goals, not every click,” says Casey Kim, an associate professor of marketing at Bryant University who studies consumer behavior and emerging technologies.
Kim expects some consumers to feel that it is unfair to hold them responsible for their agents’ mistakes. However, he does not believe the policy changes by financial institutions will deter early adopters who are already deeply invested in the technology and use these tools as part of their daily lives.
Source: www.wired.com












