Chatbots Do Not Perform Well for Misinformation-Prone Health Topics

Summary: AI chatbots might seem smart, but you shouldn't trust them with your health questions. A new study tested five popular chatbots—including ChatGPT, Gemini, and Grok—on tricky health topics like cancer, vaccines, stem cells, and nutrition.

The results? Almost half of their answers had problems. The bots did okay when talking about cancer and vaccines, but they failed badly when asked about nutrition and sports performance. The chatbot Grok gave the most "highly problematic" advice. Worse, the bots often made up fake sources and used big, college-level words that are hard to read. Experts warn that chatbots don't actually understand medicine; they just guess the next word in a sentence, which can make bad advice sound very believable.

For more details, see rssapp-infectiousdiseaseadvisor-com at infectiousdiseaseadvisor.com/news/chatbots-do-not-perform-well-for-misinformation-prone-health-topics/ (opens in new tab)