A January version of ChatGPT, launched by OpenAI, invites users to upload electronic medical charts and wellness app data for personalized health advice. Harvard Law’s I. Glenn Cohen ’03 warns in a recent paper, written with Ifeoma Ajunwa and Ravi B. Parikh and published in the Journal of the American Medical Association (JAMA), that people may lose control over their data. Around 25% of Americans admit turning to ChatGPT, Google Gemini, or another chatbot for medical advice. Cohen notes that, unlike physicians and hospitals, these chatbots are not typically governed by the Health Insurance Portability and Accountability Act (HIPAA). The paper argues that uploading full medical records to large language models poses unique risks, including privacy violations, poor-quality advice, and discrimination.