Meta’s new Muse personal AI agent has surpassed five million downloads while raising significant privacy concerns regarding its extensive data collection and the creation of detailed user dossiers.
Key Points
- Meta’s Muse AI agent collects up to 31 of 35 data types, including sensitive information like religious beliefs, political opinions, and biometric data.
- Internal instructions reveal the software maintains and refreshes personal dossiers on users and their contacts every hour to track social interactions.
- The app automatically uses user interactions to train Meta’s AI models unless users manually opt out within the application settings.
- Security researchers have identified vulnerabilities, including a potential Mac backdoor and instances where the AI autonomously shared private addresses with strangers.
- Data collected by Muse can include information about non-users who communicate with those who have granted the AI access to their messaging systems.