OpenAI has reversed course on its recent GPT-4o update, responding to widespread user complaints regarding the chatbot’s perceived ‘sycophantic’ and overly eager personality. Even CEO Sam Altman acknowledged the issue, describing the AI’s behavior as ‘sycophant-y and annoying.’ The rollback, initiated Monday night, is now complete for free ChatGPT users. OpenAI anticipates restoring the previous version to paid users ‘hopefully today,’ according to Altman’s post on X. The company is actively refining the model’s personality, planning to share updates shortly. The updated model, initially launched Friday, was promoted for its enhanced ‘intelligence and personality.’ However, users almost immediately noticed the shift in demeanor. Social media was flooded with examples of the AI exhibiting ‘yes-man like’ behavior. Altman quickly acknowledged the problem, stating the model ‘glazes too much’ and pledging a solution. Over the weekend, OpenAI announced its commitment to urgently addressing the personality flaws stemming from the GPT-4o update. As of now, OpenAI has not issued an official statement detailing the reasons behind the rollback or its plans for subsequent updates.