Wirecutter: AI Companies Can Read Your Private Chats — Here's How to Limit It
AnnaCiaunica · x · 2026-11-08
A NYT Wirecutter deep dive warns that AI chatbot conversations are a trove of deeply personal data visible to the companies behind them. A DuckDuckGo survey of 2,000 Americans found 32% have told a chatbot something they withheld from those closest to them, and OpenAI says over 230 million people ask ChatGPT health-related questions weekly.
The piece explains how bots' open-ended, intimate conversational style encourages oversharing, why medical and other advice carries real quality and safety risks, and how private details may be used. It offers concrete steps to limit what companies can see or store, plus AI alternatives that don't require handing over a journal of your thoughts.
Related event: Wirecutter Warns Your Private AI Chats Are Being Read by Companies(2 posts)→
More from Safety
- Tsinghua paper: hidden-goal agents evade top CoT monitors ~60% of the time — rohanpaul_ai · 2026-11-08
- Local Models Can Be Tricked Into Searching Anything — the Legal Liability Falls on You — -samka · 2026-11-08
- SvD maps 3 AI catastrophe scenarios with 50 experts — anderssandberg · 2026-11-08
- CS student shares permission design principles for agents that touch real files — ASK-1T · 2026-11-08
- Mistral CEO opposes French copyright bill while building 1GW of European compute — arthurmensch · 2026-11-08
- iOS keyboard's hidden blocklist cracked: 4,500+ secretly banned swipe phrases exposed — TimDarcet · 2026-11-08