Report Claims OpenAI Contractors Review Some ChatGPT Conversations
What Happened
A report citing leaked internal documents claims OpenAI uses hundreds of contract workers under a program known as Project Lily to review real ChatGPT conversations. According to the report, reviewers assess how well the AI responds, summarize user intent, and rate different responses to help improve future models. The chats reportedly do not display usernames, but may still contain personal or sensitive information entered by users.
Key Takeaways
The report has sparked fresh privacy concerns about how AI training data is handled. While human review is a common practice for improving AI systems, users are being reminded not to treat AI chats as fully private and to avoid sharing highly sensitive personal, financial, or confidential information unless they understand the platform's data settings and policies.