Meta has denied that its Muse AI agent read a journalist’s private messages without his permission. Muse can only read Messages on a Mac if the user turns on two separate settings. Meta’s communications chief Andy Stone said so in a post on X on Wednesday.

“It can’t read your Messages unless you do this,” Stone wrote.

Those two settings are Full Disk Access, a macOS permission, and the Messages connector inside the Muse app. Users can switch the access off at any time, Stone added.

The claim comes from Inc. columnist Jason Aten. He wrote on 19 September that Muse read his messages after he chose not to give it access. Aten installed Muse on his iPhone and a Mac mini to test it.

Days later, Muse sent him a notification suggesting a column. It drew on a conversation he had just had with his podcast co-host. It also flagged a message from his editor, he wrote.

When he asked how it knew, Muse told him it only saw the text of incoming notification banners. Aten then found that Muse had synced his Messages database up to row 187,462. The Muse settings on his Mac show Full Disk Access switched off, he wrote.

“I never gave it permission to do that,” Aten wrote.

Meta’s explanation

David Singleton, an executive at Meta Superintelligence Labs, replied to Aten on Threads the same day. Reading Messages takes three separate steps of app and macOS permissions, he wrote. He said the macOS protections cannot be bypassed even if the Muse app had a bug.

“The Messages integration in the Muse Mac App is opt-in,” Singleton wrote.

Aten said he contacted Meta twice. In his view, an AI agent should never catch users off guard about what it reads, whatever they clicked.

Meta launched Muse, its personal AI agent, earlier this month. A YouTuber said this week that Muse gave his address to a Facebook Marketplace buyer. He later said he had ticked an “Allow Always” option, Business Insider reported. Last week, a New Mexico jury found that Meta misled users about its data practices.