Inside 'Project Lily': OpenAI hires contractors to read real ChatGPT chats
RebeccaBellan · x · 2026-09-15
404 Media reports, based on leaked internal documents and real prompts, that OpenAI's 'Project Lily' employs hundreds of contractors to read full real-user ChatGPT conversations and rate model replies — chats that can contain highly personal information and that most of ChatGPT's 900M+ users don't know may be reviewed by humans.
- Contractors don't see usernames, but conversations often include therapist-like, professional, or intimate details
- Internal docs show the goal includes training ChatGPT to be less anthropomorphic and less sycophantic — the overly sycophantic 4o model has been cited in lawsuits tied to multiple suicides
- Reporter Joseph Cox says he has seen the review system docs and real user prompts
More from Models
- Gemini 4 Carbon vs Opus 5.5 hands-on: 3x cheaper at $4.10 and twice as fast — vitaliychiley · 2026-10-31
- 4B model Spark-X2.6 scores 94.6/100 on community benchmark, beating 27B-35B models — mesolimbic-bliss · 2026-10-31
- opencode: $100 of usage gets you DeepSeek V4.1 Flash — intellectronica · 2026-10-31
- Researcher Finds 600+ Bugs in LLM Kernel Code; Claude Blocks Safety Analysis — pmddomingos · 2026-10-31
- Signal65: Snowflake's CoCo agent finishes workflow 4.6x faster than manual — ryanshrout · 2026-10-31
- Mistral Large 4 matches GLM on coding with 47% fewer tokens, halves preview price — petrusenko_max · 2026-10-31