AI Chatbots Prone to Spreading Medical Misinformation, Studies Warn

Key Intelligence

• ■Multiple studies, including research from the University of Oxford and a Lancet study, highlight significant risks associated with AI chatbots providing medical advice.
• ■Large language models (LLMs) are not immune to medical misinformation and can even propagate it if presented professionally.
• ■Research indicates AI models are more likely to accept medical falsehoods when phrased with a professional tone, making detection difficult.
• ■Examples exist where AI-powered platforms, like Elon Musk's Grok on realfood.gov, provide nutrition information that contradicts established government guidelines.
• ■While AI can potentially learn to discern misinformation, its current susceptibility poses a public health risk.