OpenAI admits de-identified user data 'may have helped improve our models'
teortaxesTex · x · 2026-09-09
An OpenAI public statement noting "we cannot rule out that de-identified data derived from their usage of our products helped improve our models" drew mockery on X. The original quoter called it "props for straight coming clean," then digressed into personal history and authorship grievances with people at OpenAI. The reblogger joked it's "competitive" for worst public statement ever.
Related event: OpenAI Admits De-identified User Data May Train Its Models(2 posts)→
More from Fun
- Anthropic Researcher Puts AI Extinction Risk Above 10% This Decade, Admits No Alignment Plan — SIGKITTEN · 2026-09-09
- GPT 6 'Made GTA 6 Before GTA 6': 90-Hour AI-Generated Demo Goes Viral — ChrisGPT · 2026-09-09
- Satire Skewers AI Labs: 'Models Might Kill You' but First Fund the $1T IPO — zetalyrae · 2026-09-09
- Meme Take: The Archons Were the First AI Safetyists — SydSteyerhart · 2026-09-09
- Developer Builds a 'GTA 6' Demo with GPT 6 in 90 Hours — ChrisGPT · 2026-09-09
- Joke prediction: Anthropic has figured out catgirls, announcement before IPO — dejavucoder · 2026-09-09