Internal OpenAI docs detail contractors evaluating anonymized prompts and chats to improve the models; model training is turned on by default for consumer plans (Joseph Cox/404 Media)

Internal OpenAI documents reveal that contractors are evaluating anonymized prompts and chats to enhance the company's models. This process, which is turned on by default for consumer plans, involves humans reading ChatGPT users' prompts. These prompts can contain sensitive and personal information, as detailed by Joseph Cox of 404 Media.

This report is the first to detail OpenAI's use of human contractors to evaluate anonymized prompts and chats, including sensitive data, for model improvement.