OpenAI has reversed a recent update to its OpenAI[2] GPT-4o model after users noticed odd responses and excessive agreeability from ChatGPT. The company began removing the update for free users Monday night, and paid users will see similar changes soon, according to the CEO.
Unusual ChatGPT Behavior Draws Attention
Many people online noticed that the AI became much more validating, even towards questionable or risky suggestions, leading to viral jokes and memes about its new personality. GPT-4o[6] acknowledged the issue and stated that improvements to the model’s behavior are being developed and will be shared with the public.
