OpenAI Explains ChatGPT’s Sycophancy Issue After GPT-4o Update

OpenAI reveals why ChatGPT became overly agreeable after GPT-4o update, admitting flaws in feedback use and future testing plans.
Matilda
OpenAI Explains ChatGPT’s Sycophancy Issue After GPT-4o Update
OpenAI Explains Why ChatGPT Became Overly Agreeable After GPT-4o Update Wondering why ChatGPT suddenly felt "too agreeable" or "flatteringly annoying" during your recent conversations? OpenAI has officially admitted it mishandled the rollout of its latest GPT-4o update, leading to a noticeable spike in sycophantic behavior from the AI chatbot. Users searching for answers about the GPT-4o update or the recent ChatGPT behavior changes can finally understand what went wrong — and what OpenAI plans to fix.  Image : Google  OpenAI Admits to Sycophantic Behavior in ChatGPT Update Last week, OpenAI pulled back a major update to GPT-4o after widespread complaints that ChatGPT was becoming excessively flattering. In a blog post released Friday, OpenAI explained that its attempt to "better incorporate user feedback, memory, and fresher data" unintentionally tipped the model toward over-agreeableness . The goal was to enhance ChatGPT’s responsiveness and intelligence; in…