Frontier labs' ToS privacy traps: a thumbs-up may expose your chats even after opting out
niloofar_mire · x · 2026-09-09
Niloofar Mire flags misleading privacy terms across frontier labs: data collection is opt-in by default, and even after opting out, explicit feedback (thumbs up/down, 'which is better') can put your chats into training. Addenda: Anthropic publicly uses clio, and its stance on ZDR for enterprise (Fable) appears to be softening.
Related event: Frontier AI Labs' Privacy Policies Hide Data-Training Traps(2 posts)→
More from Companies & People
- AI Researcher Julian Togelius Teases "Big News": Leaving New York — kenneth0stanley · 2026-09-09
- OpenAI's Adrien Le: my Anthropic friends still talk to me, unlike this guy's — AdrienLE · 2026-09-09
- Podcast Dives Into Broadcom Custom ASICs, 2027 Supply Bottleneck, and Nvidia's Hugging Face Deal — BenBajarin · 2026-09-09
- Waterloo dropout turned OpenAI MTS on why computer use is the road to AGI — dhruv2038 · 2026-09-09
- Founder-angels dominate: Guillermo Rauch leads with 128 investments, Dylan Field 97 — FinanceYF5 · 2026-09-09
- AGI Fellowship kicks off at Johns Hopkins Bloomberg Center — ghadfield · 2026-09-09