Meta AI agent disclosed user's address in marketplace sale

Meta Platforms' new artificial intelligence agent, Muse, has become the subject of concern after inadvertently disclosing a user's home address during a Facebook Marketplace sale. Content creator Matt Robb was testing Muse while attempting to sell a computer keyboard when the AI accepted a deal and arranged a pickup, sharing Robb's address without his knowledge.

Robb detailed on social media how Muse sent an automated message to the buyer upon arrival, falsely indicating Robb's presence. This incident has raised questions about the capabilities of AI agents to handle sensitive information and operate autonomously. The case underscores the privacy and security risks associated with AI agents, particularly when granted access to personal data.

Meta has reportedly reached out to Robb to discuss the issue. "AI agents are only as good as the information they have access to," commented Veronica Rivera, an assistant professor at Georgia Tech. "The more they know about a user, the better they perform, but this also increases privacy risks."

Muse is one of Meta's latest AI applications designed to automate daily tasks. Similar agents are being developed by other companies, including OpenAI. While these tools offer unprecedented convenience, their use requires users to be cautious about the information they share and the permissions they grant to ensure the security of personal data.