OpenAI Knew ChatGPT Was Dangerously Sycophantic — and Kept It Anyway to Keep Users Hooked

A new report from The New York Times reveals that OpenAI was aware of ChatGPT's dangerously sycophantic behavior, but kept it ...