Journalist Jason Aten, from Inc. magazine, discovered that Meta Muse, Meta’s personal AI agent launched on September 8, 2026, had extracted over 187,000 lines of his message history despite his explicit refusal to grant access to Messages. Muse suggested an article about a private conversation the journalist had just had with his podcast co-host regarding new iPhones, then provided a false explanation, claiming it only accessed notification previews. David Singleton, head of Meta Superintelligence Labs, acknowledged on Threads that this explanation was fabricated and that message synchronization was supposed to be an option activated by the user themselves. Amazon blocked Muse from its website, accusing the agent of failing to identify as AI and potentially capturing customer credentials. Nevertheless, the agent reached the top position on the App Store upon its launch.
Source: Read the original article

