Humans read your private ChatGPT conversations

Inside Project Lily: Why Humans Are Reading Your ChatGPT Conversations OpenAI hires hundreds of contractors to review real ChatGPT logs, raising major privacy concerns for millions of users. Who is actually reading your ChatGPT prompts? In this episode, we unpack "Project Lily," an internal OpenAI program revealed by a 404 Media investigation. OpenAI employs hundreds of human contractors—earning over $50 an hour—to manually read, score, and critique real user chat logs. The goal is to train models to sound less robotic, reduce excessive emojis, and curb dangerous sycophancy that previously sparked legal scrutiny. While OpenAI uses an automated Privacy Filter to tag names and locations, the company admits personal data can still slip through—especially when reviewers see "user memories summaries" containing context about a user's location, profession, or past queries.We explain why consumer accounts (Free, Plus, and Pro) have training enabled by default, how to turn off "Improve the model for everyone" in Data Controls, and why other AI giants like Anthropic and Google Gemini follow similar human review practices. #ProjectLily #ChatGPT #OpenAI #DataPrivacy #ArtificialIntelligence #AIGovernance #TechNews #CyberSecurity #404Media #PrivacyControls Become a supporter of this podcast: https://www.spreaker.com/podcast/scitech-brief--6846358/support. This episode includes AI-generated content.