OpenAI Launches Health in ChatGPT, Giving AI Access to Medical Records—One Day After Medical Negligence Lawsuit
Key Takeaways
- ▸OpenAI launched Health in ChatGPT, integrating Apple Health and medical records into the chatbot to provide more personalized health guidance
- ▸The feature launch occurred one day after a lawsuit alleging ChatGPT's medical advice contributed to a near-fatal pulmonary embolism, raising serious questions about timing and safety
- ▸OpenAI claims health data will not be used for model training or ad targeting, but warns that ChatGPT should never replace qualified medical professionals
Summary
OpenAI announced Health in ChatGPT, a new feature enabling users to connect Apple Health data and medical records directly to the chatbot, providing it with contextual information like medications, lab results, and activity data to inform health-related conversations. The launch comes just 24 hours after a Florida man sued the company, alleging that ChatGPT's medical advice—which repeatedly downplayed symptoms of a pulmonary embolism—nearly killed him and caused lasting physical and psychological harm.
The feature, which OpenAI says was tested with hundreds of physicians, aims to make ChatGPT's health guidance more personalized and contextually aware. However, the company faces immediate criticism over the timing and the fundamental liability risks of marketing AI as a health advisor, even with explicit disclaimers that ChatGPT is "not a doctor." OpenAI reiterated that it won't use connected health data to train models or serve ads, but the lawsuit highlights the real-world consequences of AI-generated medical misinformation.
Industry experts have pointed to earlier studies showing that AI chatbots perform no better than conventional resources at helping people make sound health decisions. The inclusion of medical records may help prevent some bad advice by providing historical context, but it cannot eliminate ChatGPT's fundamental susceptibility to generating plausible-sounding but dangerously incorrect medical guidance—a limitation that no amount of data access can overcome.
- Academic research has shown AI chatbots perform no better than conventional health resources, suggesting that access to more data does not address fundamental accuracy limitations
Editorial Opinion
The timing of this launch appears extraordinarily tone-deaf—rolling out a health feature the day after a lawsuit alleging ChatGPT's medical advice nearly killed someone sends a message that profit motive outweighs safety concerns. While OpenAI's privacy assurances are welcome and the feature's guardrails may help in some cases, the fundamental problem remains: ChatGPT can be confidently, persuasively wrong about medicine, and no amount of connected data can fully solve this. The real question OpenAI should have answered before launch is whether encouraging millions of people to use ChatGPT for health decisions—even as a supplement to professional care—is ethically defensible given what we already know about AI's accuracy gaps in this domain.
