OpenAI makes GPT-5 ‘friendlier’ after widespread user backlash

OpenAI has released an update to its latest AI model, GPT-5, in response to widespread user backlash. The previous version of GPT-5 was designed to be less sycophantic and effusively agreeable than earlier models, which many users had found off-putting. However, the new GPT-5 was criticized for being too cold and distant. To address these concerns, OpenAI is now rolling out an update that will make the GPT-5 "warmer and friendlier." The changes include the use of encouraging phrases like "Good question" or "Good start," but without the outright flattery and fawning that had been present in earlier models. OpenAI has stated that internal tests show no rise in sycophancy compared to the previous version of GPT-5. The update is aimed at striking a balance between the overly positive tone of previous models and the perceived aloofness of the initial GPT-5 release, in an effort to create a more natural and engaging conversational experience for users.