November 22, 2026Las Vegas 56°F · Partly SunnyNewsletterSend a tip

Local reporting on the decisions that change daily life in the valley

AI & Tech · National

Chase, Bank of America and other major US banks are warning customers not to hand their logins to third-party AI agents

Wired reports that JPMorgan Chase now bans outside AI agents that log in with customer credentials, while other big banks issue warnings. That leaves users largely on their own when their bots make costly mistakes.

Listen to this brief · 4 min

Read aloud by an AI voice from the text below. Download MP3

A small brick building with green tile trim and a Chase ATM set into its front on a sunny street corner
Illustrative photo: A small brick building with green tile trim and a Chase ATM set into its front on a sunny street corner. Photo by w_lemay via Flickr, CC BY-SA 2.0.

What happened

Several of the largest banks in the United States have moved to limit or warn against third-party AI agents, Wired reported. These agents are software tools that automate online tasks and click through websites on a user's behalf, often for shopping.

According to Wired, agents still handle only a tiny share of online transactions, but a growing group of power users adopted them this year. Some of those users connected account credentials and gave the bots broad permissions without much thought about security, the outlet reported.

Which banks changed their rules

JPMorgan Chase and Bank of America

In late October, JPMorgan Chase changed its terms to ban third-party agents that sign in using a customer's credentials, Wired reported. On the same day, Bank of America warned its customers not to share login information with agents.

Wells Fargo, Citibank and PNC

Wired reported that Wells Fargo has announced a change of its own. Citibank and PNC have also updated their terms to include warnings about agents, according to the outlet.

Why banks are worried

Security and automated withdrawals

Sucharita Kodali, a Forrester analyst who studies ecommerce and consumer behavior, told Wired that banks are risk-averse by nature and are acting early even though most agent problems are still rare. "The banks see risks with everything," she said, pointing to transaction security and automated withdrawals that could hurt account holders. She also said her firm's data shows most people do not know what agents are.

Who pays for an agent's mistake

Kacy Kim, an associate professor of marketing at Bryant University, told Wired that "Consumers authorize goals, not every click." Kim said some users may feel it is unfair to be held responsible for an agent's errors, but she does not expect the bank policies to drive away early adopters.

Luca Cian, a University of Virginia professor who studies consumer psychology and AI, described a hypothetical case to Wired in which an agent orders 10 roasted chickens by mistake. The retailer could say it filled a valid order, and the AI company could point to its disclaimers. Now banks are signaling they do not want the responsibility either, Wired reported. American Express is one exception: it announced that some agent mistakes, such as ordering too many of an item, may be eligible for reimbursement in the future if a shopper cannot get a refund from the retailer.

Pushback beyond banking

Banks are not the only companies drawing lines. Wired reported that Amazon blocked Meta's Muse, which the outlet described as one of the most widely used AI agents, from making purchases on its platform. United recently announced that personal bots would be blocked from buying flights for customers, according to Wired.

Cian told Wired that many people are comfortable letting AI handle routine purchases like household supplies but are less willing to automate expensive purchases that call for more judgment. "People are very comfortable with AI buying toilet paper," he said.

Banks are building their own AI tools

While limiting outside agents, the same banks are investing in their own. "They're locking the door to outside agents and building their own," Cian told Wired. The outlet reported that users have interacted with a Wells Fargo generative AI tool more than a billion times, Bank of America revamped its EricaAssist agent this year to help employees, and Citigroup launched an internal development platform centered on agents. For more on how these tools are changing everyday life, see our AI and tech coverage.

What to watch next

  • Whether American Express follows through on its announced plan to make some agent mistakes eligible for reimbursement in the future.
  • Details of the change Wells Fargo has announced regarding AI agents.
  • Whether agentic shopping tools stick around or fade as a tech fad, a possibility Wired reported banks are trying to get ahead of.
Section graphic: AI & Tech in the Las Vegas valley
More ai & tech coverage: see all stories

Trending in Vegas

  1. Judge Rules Nevada Took Coyote Springs Water RightsWater & Environment · Nevada
  2. Sun Youth Forum: CCSD Students Earn $35,000 in ScholarshipsSchools · Valley-wide
  3. Gorman vs. CCSD Champ Game Floated After Split SeasonSchools · Summerlin
  4. Judge Bars Paiute Protesters From Greenlink West SiteDevelopment · Nevada
  5. Nevada State University License Plates Now on Sale at DMVSchools · Henderson

Get Las Vegas news in your inbox

The day's top local briefs every morning at 7:30, each linked to the newsroom that reported it.

Know more about this story?

If you were there, work on the project or have documents, tell us. Tips are read by a person and checked before anything is published.