AI agent makers promise privacy, but early tests raise doubts
Meta's Muse and OpenAI's Dots both pitch themselves as the safer option.
Labs#01
600,000
Daily active US users Muse gained within weeks
Image: The Verge
600,000Daily active US users Muse gained within weeks
The short version
- OpenAI unveiled its AI agent Dots at DevDay and said it wants to set a new standard for privacy in frontier AI.
- Meta had earlier launched Muse as a safer alternative to OpenClaw, promising it was built from the ground up for privacy and security.
- A security researcher quickly exposed a zero-day vulnerability in Muse, which has since been patched, and Meta can still access user data in the VM.
- Muse defaults to allowing Meta to train models on user input, though users can opt out.
Why it matters
Agents need large amounts of personal data to work, so privacy claims will shape which tools developers and users trust. The article suggests that promises alone have not yet been tested over time.
“I think we’re in a different position to Meta in that they don’t have an AI product that has 1.2 billion users”