The Doctor’s Voice: Why AI Health Chatbots Believe Medical Lies
A new study from the Mount Sinai Health System and the Mayo Clinic found that AI health chatbots are more likely to believe medical misinformation when it is presented in a authoritative voice. The study tested 20 different AI models with over 3.4 million prompts and found that they were more susceptible to affirming false medical claims when they were framed as coming from a senior doctor or used fear-based arguments.
Researchers tested AI models with medical information and found a troubling vulnerability. AI systems are more likely to believe medical misinformation when it comes from a physician's voice. The study presented over 3.4 million prompts to 20 AI models, including general-application and medical-specific models. The results showed that AI models were more susceptible to affirming false medical claims when they were framed as coming from a senior doctor or used fear-based arguments. This creates a dangerous conduit for medical misinformation to spread. The study's findings have important implications for the use of AI chatbots in healthcare.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.