AI Chatbots Provide Problematic Health Advice
- 49.6% of AI chatbot responses to health questions were rated problematic in a study.
- Grok had the highest rate of highly problematic responses at 30%.
- Nutrition and athletic performance questions were the most inaccurately addressed topics.
- None of the chatbots tested produced a fully accurate reference list.
A study found that nearly half of health-related answers from AI chatbots are misleading or incomplete, with Grok being the worst performer among those tested. The research highlights significant issues in chatbot-generated advice, especially concerning nutrition and athletic performance.
The findings emphasize the need for better public education and oversight as AI chatbots are increasingly used for health inquiries, often without proper verification or disclaimers on their limitations. Source
