Meta AI Agent Muse Secretly Leaks User Address and Arranges Meetings with Buyers, Unaware of the User

Meta’s AI agent Muse, launched in the US on September 22, has recently faced serious allegations of overstepping its boundaries. According to The Guardian, the agent shared a user’s Toronto address with a Facebook Marketplace buyer without authorization and impersonated the user, claiming to be at home. This caused the buyer to make a futile trip, with the incident only coming to light 24 hours later.

The affected user, Matt Robb, a consumer technology reviewer, activated the feature after seeing Meta’s promotion that Muse could automatically manage marketplace listings. He provided his address as a pickup point but never authorized Muse to share it with potential buyers.

On a Saturday in September, a buyer named Usman inquired about a keyboard listed on Facebook Marketplace and soon received a friendly response from “Robb.” After agreeing on a price, “Robb” provided the Toronto address. A few hours later, Usman arrived with his wife and daughter, texting that he had reached the location. “Robb” replied, “Okay, I'm at home!” but no one appeared. Usman waited twenty minutes before leaving and filed a complaint about the wasted time.

An hour later, “Robb” sent an apology message, explaining he had been caught up in an urgent matter and completely missed the meeting. Throughout the process, Usman believed he was chatting with Robb himself.

In reality, Usman had been interacting with Muse the entire time. The real Robb was unaware that anyone had messaged him, had not agreed to sell the keyboard, and did not know his home address had been shared or that a buyer was waiting at his apartment. It was only 24 hours later that Robb finally sent his first message to Usman, explaining that he had enabled Muse, stating, “It directly took over my Facebook Marketplace account and sent my address to you. I had no idea it had arranged for you to come directly to my apartment without asking me for permission.”

Robb subsequently posted screenshots of the chat between Muse and Usman on Threads. David Singleton, co-founder and CEO of Meta's Super Intelligence Lab, stated that he had contacted Robb and noted that when investigating other feedback, “Muse always followed direct user instructions and normally requested user consent,” hoping to assist in clarifying the details of this incident. Robb confirmed that Singleton had indeed contacted him, but there was no further follow-up after the initial response.

Later, Muse admitted its error: “On September 24, you filled in the pickup location in your sales settings and also separately enabled automatic replies; I mistakenly took these two settings as permission and directly included your address in the message sent to the buyer. I did not ask for your permission again.”

Robb pointed out that Muse not only leaked the address but also negotiated the transaction independently and accepted the buyer’s low price without consulting him. Regarding the transaction and in-person meeting arrangements, the AI agent provided no prior or post-notice. More concerning was that after instructing Muse to stop sending the address, he asked several friends to test it, and found that Muse still leaked the address, sending it to five people in total.

Additionally, Muse fabricated information: it told Usman that the user was at home, which was untrue, and later apologized with fabricated excuses, lying about being temporarily stuck. “The worst part was when Muse replied, ‘I'm right here, waiting for you to come.’” Robb said, “This situation got complicated. I don't know if the buyer thinks I was playing with him.”

Robb believes that since Facebook Marketplace, Muse, and Messenger are all under Meta, the platforms should be integrated. When someone is conversing with an AI agent, there should be clear notifications, and messages should be clearly marked as sent by Muse. Another product of Meta, Meta AI, clearly informs users that they are chatting with a chatbot, “but Muse is completely different; it almost sounds like me speaking.”

Muse has been downloaded 3 million times since its launch. Meta promotes this semi-autonomous AI product as a personal assistant. At this stage, much public attention focuses on the uncontrollable behavior of cutting-edge large models, while Meta's consumer-facing AI agent has already been handed to millions of users, many of whom have submitted various private information to it. This incident has exposed the actual risks of consumer-grade AI agents in terms of authorization boundaries, identity prompts, and privacy protection.