ChatGPT user feedback has prompted OpenAI to roll back the latest GPT-4 update, due to the chatbot’s annoyance and sycophantic personality developed with the update.
Developed with improved “intelligence and personality” the change was recognised as soon as it was implemented, with users commenting on the yes-man nature of the chatbot’s responses.
CEO Sam Altman quickly acknowledged the issue on Twitter, and promised a rollback of the update alongside fixes to the model to address the change in personality; the rollback is expected to be complete for paid users soon.
Further updates are promised from OpenAI in the coming days.
The rapid reaction to customer feedback is in contrast to the slow response to violent content generating from GPT-3, and is an encouraging sign for the future of the product.