OpenAI says ChatGPT outperforms doctors, but I’m not buying it

LLM chatbots may have correct medical information in their training database, but a new study shows that people who use them for medical advice end up with wrong answers.