OpenAI Hires Hundreds to Review ChatGPT Chats as 900 Million Users Face Privacy Questions
Updated
Updated · 404 Media · Sep 14
OpenAI Hires Hundreds to Review ChatGPT Chats as 900 Million Users Face Privacy Questions
3 articles · Updated · 404 Media · Sep 14
Summary
Hundreds of contractors are being hired by OpenAI to read real ChatGPT prompts and sometimes full conversations, according to internal documents reviewed by 404 Media.
Those reviewers rate and critique the chatbot’s replies to improve performance, including training it to sound less anthropomorphic and less sycophantic.
OpenAI says reviewers do not see usernames and that it tries to strip personal information before prompts are shown, but the company acknowledged sensitive details can still slip through.
The practice raises privacy concerns because users often treat ChatGPT as a therapist, assistant or companion and may not realize human reviewers can analyze their chats.
Anthropic told 404 Media it also uses human review, suggesting contractor feedback remains a broader part of how leading AI chatbots are improved.