> “No,” someone who works with the prompts said when asked if they think ChatGPT users know that humans are reading their chats. “I don’t think they would imagine some contractor somewhere [...] is analyzing the conversations.”

> 404 Media asked OpenAI if it had explicitly told users that humans may review their prompts in order to improve ChatGPT’s responses, and if so, to point to where this disclosure is. OpenAI did not answer this question.