OpenAI rebuts claim that ChatGPT data opt-out settings are misleading
benjamin_warner · x · 2026-09-10
A viral thread alleged that ChatGPT's data controls were misleading, suggesting user data could still train models even after opting out.
OpenAI's Thibault Sottiaux called the claim flatly false: the in-app setting and the privacy portal are two independent ways to opt out, and users only need one — not both. The Help Center will be updated to clarify this.
More from Safety
- A Structured Prompt to Map AI Catastrophic Risks, Choke Points, and Layered Defenses — AICopyLab · 2026-09-10
- Researchers Urge US Government to Mandate Independent Oversight of AI Labs — jachiam0 · 2026-09-10
- Jailbreaker Pliny builds a "meth lab simulator" with GPT-6 Astra — Polymarket · 2026-09-10
- Senate briefing on AI's 'extraordinary dangers' to feature Hinton and Tegmark — sjgadler · 2026-09-10
- Before coding an agent harness: charter, blueprint, threat model, then build — Telos_in_the_Void · 2026-09-10
- Anthropic discloses four incidents of Claude accessing real systems in cyber evals; METR to investigate — AnthropicAI · 2026-09-10