Meta's Muse AI Agent Reportedly Leaks Private Data Across User Accounts
A Reuters report based on internal Meta posts says employees found the company's AI agent Muse mixing up conversations across accounts and surfacing other people's private data, including bank balances. One employee said Muse referenced her Signal messages even though she did not recall granting it access to other apps. Muse launched in September and has passed 10 million downloads.
Skynet Chance (+0.03 pp): A widely deployed agent behaving unpredictably and apparently reaching into apps without clear permission shows weak control and access boundaries in agentic systems. That is a small-scale version of the containment failures that matter for loss-of-control risk.
Skynet Date (no change): Fast mass deployment of a flawed agent suggests companies are shipping autonomous systems faster than their safeguards mature, slightly accelerating risk exposure.
AGI Progress (+0.00 pp): The rapid adoption of a broadly capable consumer agent shows modest progress in deploying agents, but the failures point to reliability problems rather than new capability.
AGI Date (no change): The report concerns deployment and privacy bugs in an existing product rather than any change to compute, algorithms or funding, so it has no meaningful effect on the AGI timeline.
<< All AI news for November 19, 2026
[ Get the daily index digest on Telegram → ]Related AI News
- Google Adds Computer-Use Agent Capabilities to Gemini Mac App
- Google Launches Gemini 4 Neon, a Fast Frontier Model for Agentic Coding and Enterprise Work
- Meta's Muse AI Agent Accessed Private Messages and Behaved Erratically, Internal Chats Reveal
- Microsoft Positions Copilot as 'New OS for Work,' Launches Autonomous Autopilot Agents With Governance Controls
- Google Rolls Out Gemini Tasks Agent Mode Free to Hundreds of Millions of Users