OpenAI has withdrawn the latest update to its GPT-4o model, which is used by default in ChatGPT, after users complained about the chatbot’s overly flattering responses and tendency to agree with everything. The company’s CEO Sam Altman announced the withdrawal of the update in a post on X (formerly Twitter), TechCrunch reports.
"We started rolling back the latest update to GPT-4o last night," Altman wrote. "It’s now 100% rolled back for free [ChatGPT] users and we’ll update again when it’s finished for paid users, hopefully later today. We’re working on additional fixes to model personality and will share more in the coming days."
The update, which was rolled out late last week, sparked a wave of social media posts, with users mocking ChatGPT's new "flattering" behavior. Many posted screenshots of the AI admiring or even endorsing questionable or dangerous ideas, which became a viral meme.
On Sunday, Altman acknowledged the issue for the first time and promised to quickly fix it and share the findings. With today's withdrawal of the latest GPT-4o update, both free CahtGPT users and paid subscribers will receive more balanced and moderated responses while the model's "personality" continues to be refined.