An audit of chatbot responses in health and medical fields prone to misinformation found that 49.6% of responses were problematic. Specifically, 30% of responses were somewhat problematic, and 19.6% ...