OpenAI’s ‘what we missed’ posting looks at flawed rollout of ChatGPT update

Generative artificial intelligence leader OpenAI has posted an examination of “what went wrong” in the rollout of its latest ChatGPT model update, which critics pounced on for being so agreeable with prompts that it would promote unsafe behavior and led company CEO Sam Altman to acknowledge it was “too sycophant-y and annoying."

The company’s latest moves appear aimed at demonstrating a healthy governance process to deal with miscues in the development of generative AI systems, as the federal government’s emphasis...