Meta faces scrutiny over alleged unauthorized access by AI agent muse

Meta is facing criticism after a journalist claimed its AI agent Muse accessed private messages without user consent. The company's VP of Communications, Andy Stone, denied the allegations, stating that the Messages integration in Muse requires explicit opt-in permissions. Users must enable Full Disk Access and the Messages connector for Muse to read their messages. Stone emphasized that the app cannot access messages without these steps. Despite Meta's denial, skepticism remains due to the company's history of data mishandling. David Singleton, from Meta Superintelligence Labs, added that the app's security features make unauthorized access nearly impossible. However, the journalist, Jason Aten, argued that Muse accessed his messages without Full Disk Access enabled, suggesting a potential breach. Singleton disputed this, claiming the AI provided an incorrect explanation. The incident highlights ongoing concerns about Muse's privacy practices, with another user reporting a similar issue involving Facebook Marketplace. Meta is investigating these claims, underscoring the challenges it faces in maintaining user trust in its AI products.