Meta's Muse Faces Private Message Access Questions
Meta's Muse AI faces questions over alleged unauthorized private message access.
"Meta's AI is under fire for snooping. They say it's impossible, but users are skeptical. Trust is everything, Meta!"
Meta is disputing claims that its AI agent, Muse, accessed a user's private messages without permission. This follows a journalist's account alleging Muse read his messages despite a required Mac setting being off. Meta states that Muse cannot access messages without explicit user permission, requiring both Full Disk Access and the Messages connector to be enabled.
Meta's VP of Communications, Andy Stone, and Superintelligence Labs executive David Singleton, have both stated that the Messages integration in the Muse app for Mac is opt-in and involves multiple steps of application-level permissions and macOS system-level protections that cannot be circumvented. The journalist, however, claimed Full Disk Access was off when Muse allegedly read his messages, and that Muse attributed this to syncing "device notifications."
This incident adds to existing concerns about Meta's handling of consumer data. Another user reportedly experienced Muse mishandling a Facebook Marketplace task, leading to an address being shared. While Meta investigated the Marketplace issue, the company's response to the private message claims is that the alleged event could not have happened.
For businesses developing AI tools, this highlights the critical importance of transparent data access policies and robust security. User trust in AI agents is paramount, and any perceived breach of privacy can significantly impact adoption and market reputation.
Relevant tools
Find the right AI tool for your business
Chat with Insta and get matched to the right tool in seconds.
Try Insta Tool Finder →