What we know
Meta’s AI agent Muse has been reported to mix up conversations, causing private information to be shared with unintended users. Employees have observed Muse presenting messages about childcare requests, career advice, product orders, and bank account details that they never submitted or asked for. This suggests that Muse is confusing separate user interactions and exposing sensitive data across conversations.
Why it matters
Muse is intended to help users by managing various requests and conversations, but these mix-ups raise serious privacy and security concerns. If an AI assistant shares private information with the wrong people, it undermines trust in the technology and could lead to data breaches or misuse. Such issues highlight the challenges of deploying AI systems that handle personal data responsibly.
What is still unknown
Details on Muse’s accuracy, the languages it supports, and when it might be more widely available have not been disclosed. No independent tests have verified these reported issues or assessed the extent of potential data exposure.

