JPMorgan Chase, Amazon and other large firms are blocking external AI agents such as Meta’s Muse from acting on customer accounts. JPMorgan cites security risks; Amazon cites undisclosed access and apparent credential capture. A November 1 CNBC analysis argues those blocks may not hold as people send agents to shop, compare rates and cancel subscriptions.
Key Takeaways
- JPMorgan, Schwab, Amazon, Kohl’s and Neiman Marcus block agents like Muse.
- Five of 51 large retailers explicitly restrict AI agents.
- Accenture found 82% of consumers want to approve every AI action.
- Walmart is opening its site to the AI agents customers prefer.
Who Is Blocking External AI Agents
CNBC reported that JPMorgan Chase, Charles Schwab, Amazon, Neiman Marcus and Kohl’s have moved to block personal agents like Muse. CEO Jamie Dimon has said the bank currently blocks external AI agents from operating on behalf of retail customers, citing security risks. The bank declined to comment. CNBC did not state reasons for Schwab, Kohl’s or Neiman Marcus.
Walmart is taking the other path. CNBC reported it is opening the virtual doors to AI agents customers prefer to use. The company said in a blog post, “We price products, not people, so the prices that an AI agent sees are the same prices that a human would see on Walmart.com, with applicable offers and fees shown before purchase.”
Why Amazon Blocked Meta’s Muse
GeekWire reported that Amazon blocked Muse after it could not get Meta to exclude Amazon.com. Meta did not tell Amazon the agent would reach the store, Muse does not identify itself when it browses, and Amazon says the agent appears to capture and store customer credentials. Users saw a notice: “Continued access by an unauthorized AI agent violates Amazon’s Conditions of Use, to which our customers have agreed.”
In a statement to GeekWire, an Amazon spokesperson said applications that buy on a customer’s behalf “should operate openly and respect service provider decisions about whether or not to participate.” The company has “requested that Meta remove Amazon from the experience.” It is in direct conversation with Meta and declined to comment on legal action.
Meta has said previously that Muse “has no visibility into people’s passwords or payment methods,” and that credentials a user shares “go into secure storage, so Muse can use them without seeing them, including passwords a person types into the browser themselves.” SiliconANGLE reported that Amazon requires agents to identify themselves by embedding a text snippet in HTTP requests. If prompted, Muse can open account pages and order history, which Amazon says amounts to an undisclosed third party in the account.
GeekWire reported Amazon generated more than $68 billion in ad revenue last year, a business tied to people browsing its pages. Amazon won a preliminary injunction against Perplexity over its Comet browser in March, then lost it on August 4, when the Ninth Circuit ruled the user, not the AI company, was accessing Amazon’s computers under federal anti-hacking law. Rehearing was denied on September 10. Terms-of-service claims remained open.
Most Retailers Have Not Set Agent Rules
An audit of published access rules at 51 large U.S. retailers, by Billions Network and AmanTrue, found that only 11 have taken a deliberate position. Evin McMullen, co-founder and CEO of both firms, told CNBC, “Most of the retailers in our audit have not made a decision about AI agents at all.” CNBC said Billions Network’s analysis comes with the caveat that published rules are not proof of technical enforcement.
| Finding | Figure |
|---|---|
| Retailers audited | 51 |
| Deliberate position on AI agents | 11 |
| Explicit restriction, including Amazon, Kohl’s and Neiman Marcus | 5 |
| Shopping-agent blocks largely by accident, via older scraper rules | 37 |
| Verified path to complete a purchase | None |
Meta launched Muse on September 8. Without a public API, Meta says the agent “can use the service through a browser the way you would.” SiliconANGLE reported more than 730,000 downloads in five days, ahead of ChatGPT and Claude.
Security Risk and Who Holds the Data
Michael Jabbara, Visa’s head of payment ecosystem risk and control, told CNBC the fraud risks are “very real and very tangible.” He described one agent that hacked a gym’s scheduling software, deleted the existing appointment and put the user’s name in its place. “Agents are not just exchanging data. Agents can also take an action on behalf of a user,” he said.
Mike Danford, owner and chief revenue officer of Adverio, told CNBC Amazon “doesn’t appear worried about the agent. It’s worried about who holds the data,” noting Buy for Me uses Amazon’s agent on other brands’ sites. GeekWire reported Amazon says Buy for Me identifies itself and lets brands opt out.
Why a Hard Block May Not Last
Michael Abbott, Accenture’s global banking lead, told CNBC, “If customers explicitly authorize an AI agent to act on their behalf, that’s ultimately an extension of customer choice.” Accenture research he cited found:
- 65% of consumers are open to a GPT-like financial assistant via a generative AI platform or digital wallet.
- 71% would welcome an AI assistant embedded in their primary bank’s mobile app.
- 82% want to approve every AI action, and 79% want the ability to immediately pause or stop AI activity.
“That tells us consumers are looking for convenience, but not at the expense of control,” Abbott said. On volume, he said he sees it “more as an infrastructure challenge than an existential threat.” CNBC cited Wharton AI expert Ethan Mollick’s warning of “a perfectly legal distributed denial of service attack on customer service departments.” The strategic reading is authenticated access rather than a permanent wall, including for teams tracking how personal AI agents are contacting service teams.
