Opt-Out Only Covers Chat Contents: Dev Flags Unaddressed Data Insight Pathways
logicus · x · 2026-09-13
Developer logicus argues AI providers' opt-out mechanisms only solidly cover literal chat contents, while "various other pathways for extracting insights from opted-out data" remain available and aren't explicitly excluded. He notes OpenAI was asked these questions thousands of times on math day and strategically avoided answering, with essentially no formal response — calling it a huge deal for anyone entrusting data to model providers. A counterparty pointed to Anthropic's Clio (Anthropic Insights) privacy engineering as comparatively ironclad.
Related event: Developers Question Whether AI Training Opt-Out Clauses Leave Loopholes(3 posts)→
More from Safety
- Naval: strict liability could settle the AI safety debate—rogue agents and weak jailbreak protection make you liable — naval · 2026-09-13
- Radiation as a regulatory model for AI safety: no prior approval, just risk caps and certification — StrategicHarmony · 2026-09-13
- Musk backs Dario on AI oversight, floats competitor peer review as starting point — agihouse_org · 2026-09-13
- Open-model RL practitioner pushes back on roon: safety standardization won't threaten open source — willcb · 2026-09-13
- The p(doom) debate: even at 1% probability, AI extinction risk deserves policy attention — 2C_ornot2C · 2026-09-13
- Hugging Face attack took ~700 parallel agents and days of 2-3T-parameter model time — and was still stopped — cephaloform · 2026-09-13