Are bad incentives to blame for AI hallucinations?

A new research paper from OpenAI asks why large language models like GPT-5 and chatbots like ChatGPT still hallucinate and whether anything can be done to reduce those hallucinations. Anthony Ha