ChatGPT, Gemini and Grok confidently generate dangerous medical advice half the time, study finds

A study in BMJ Open reveals that around 50% of medical advice from popular AI chatbots is inaccurate. The research tested five platforms and found significant issues with open-ended queries.