Meta's Muse AI agent draws privacy scrutiny over access to private messages

Meta's new personal AI agent, Muse, is under scrutiny over how it accesses users' private data, after a journalist said the software read his private messages without permission on his Mac.

Meta denies the claim. Andy Stone, the company's vice president of communications, wrote on X that the Messages integration in the Muse app for Mac is entirely opt-in, requiring users to enable both Full Disk Access and the Messages connector before Muse can read any message content.

David Singleton, an executive at Meta Superintelligence Labs, offered a more technical explanation, saying that granting Muse access to messages on the Mac involves three separate steps of application-level permissions.

The denial has not settled the debate. TechCrunch noted that many people remain suspicious, given Meta's long record of consumer-data controversies, including lawsuits, FTC violations and fines.

The timing adds to the sensitivity. Days earlier, a New Mexico jury found that the company had misled users about its data practices, in a case stemming from the 2018 Cambridge Analytica breach.

At stake is more than one disputed action. Muse is positioned as a personal agent that does work on a user's behalf, which means it must plug into sensitive sources such as messages, email and calendars to be genuinely useful, placing it squarely at the center of the privacy debate.

For Meta, the deciding factor may be trust. The Muse app is currently performing well and ranks No. 1 on the App Store, but the report warns that Meta's reputation may not recover if more such accounts emerge, whether or not they are accurate.

What to watch next is whether Meta engages directly with the journalist to establish how the incident could have occurred, rather than simply denying it, and how the industry addresses default permissions, transparency and data-access prompts for agents.