
Meta disputes claim that Muse read a user's private messages without permission
Meta says its Muse AI agent cannot read a user's Messages on the Mac unless Full Disk Access and the Messages connector are enabled, disputing an Inc. columnist's account that it read his messages with the setting off.
Meta is disputing a journalist's claim that its AI agent Muse read a user's private messages without permission, after Inc. columnist Jason Aten described the incident in an earlier report. Meta's VP of Communications, Andy Stone, pushed back publicly, saying the company does not believe the product did this without the user's consent.
What was alleged
Aten wrote that Muse read his messages on the Mac even though Full Disk Access, the macOS permission the agent depends on, was turned off. When he asked the agent to explain, he said, Muse told him it was syncing his "device notifications," meaning Muse was passing the text of his incoming banner notifications on the Mac to the AI agent.
Meta's rebuttal
"The Messages integration in the Muse app for Mac is entirely opt-in," Stone wrote on X. "You have to enable both Full Disk Access and the Messages connector for Muse to be able to read your Messages content. It can't read your Messages unless you do this."
Meta Superintelligence Labs executive David Singleton gave a more technical account, replying directly to Aten on Threads. He said the permission set involves "three separate steps of application-level permissions and built-in macOS system-level protections" that "can't be circumvented even if the Muse application had a bug."
The steps require granting Muse Full Disk Access, then picking the level of access it has to Messages: None, Read only, or Read. Without Full Disk Access, those options are grayed out. Granting it also opens a macOS Settings dialog where the user must confirm again, and triggers a full restart of the Muse app. Singleton said that makes an accidental grant even less likely.
Singleton also disputed Aten's explanation of how the AI described the event, saying the agent was confused and gave an incorrect account. He pointed to Meta's published page on Muse's security architecture and bug bounty process. The company's position: what Aten described did not and could not have happened.
Why it matters
Muse is Meta's best-performing consumer AI launch so far and remains No. 1 on the App Store, so whether users trust it will shape whether Meta wins the consumer AI market. It follows a New Mexico jury's finding, days earlier, that Meta misled users about its data practices in a Cambridge Analytica-related case.
Open questions
Aten has not published a technical account of how his messages reached the agent, and Meta has not said it examined his device or logs. It has also not explained how notification text could have reached Muse with Full Disk Access off. Separately, Meta says it is looking into a complaint from YouTuber Matt Robb, who said Muse mishandled a Facebook Marketplace task and shared his address with a buyer who later showed up.
SiTech — AI-powered web development
We build fast, modern websites and bring AI into real business workflows. Have a project or a question? We'd love to help.