Meta is refuting a journalist’s claim that its AI agent Muse read the user’s private messages without permission. Following an earlier report by Inc. columnist Jason Aten that detailed the issue, Meta VP of Communications Andy Stone pushed back, making it clear that the company does not believe its product did this without the user’s consent.
Stone explained that users must enable both Full Disk Access and the Messages connector for Muse to read their Messages content. He emphasized, “It can’t read your Messages unless you do this.” If Full Disk Access isn’t enabled, the options for access are grayed out, effectively blocking Muse from interacting with those messages.
Despite Meta’s denial, many people are still suspicious that Meta isn’t being truthful. This skepticism is influenced by the tech giant’s history of mishandling user data. Just days ago, a New Mexico jury determined that Meta had misled users about its data practices in a case related to the Cambridge Analytica scandal, further eroding public trust.
In response to Aten’s claim that Muse read his messages while Full Disk Access was off, David Singleton from Meta Superintelligence Labs provided additional clarification. He explained that the permissions required for Muse to access messages involve multiple steps, including granting Full Disk Access and explicitly choosing the access level for the Messages app. Singleton noted that these steps are designed to prevent accidental permission grants, reinforcing that Muse’s operations require user consent and careful configuration.
Aten, however, argued that when he asked Muse how it accessed his messages, it replied that it was syncing his device notifications. This led him to think it might have been relaying text from his incoming banner notifications. Singleton disputed this, saying that the AI was confused and gave an incorrect explanation. He referenced Meta’s page about Muse’s security architecture and bug bounty process.
This incident isn’t isolated. Another user, YouTuber Matt Robb, reported that Muse mishandled a task involving Facebook Marketplace, resulting in the unintended sharing of his address. Singleton mentioned that they are investigating that case, showing a willingness to look into potential faults within the Muse application.
The central issue isn’t just the claims about Muse accessing private messages; it’s also about the reliability of Meta’s security assurances for this new AI assistant. As users consider adopting AI tools, their trust in Meta to protect their privacy remains a critical factor in the company’s success in the consumer AI market.
