Dr. ChatGPT is now officially in, as OpenAI launches dedicated health section for its chatbot

"Ask Dr. Google" used to be a running joke about patients who researched symptoms between doctor visits. AI is taking digital self-diagnosis to a whole new level, and OpenAI sees a major opportunity.

OpenAI has launched ChatGPT Health, a dedicated section within ChatGPT focused on health and wellness. The company says users can securely connect their health information to the chatbot "to help you feel more informed, prepared, and confident navigating your health."

Over 230 million people worldwide ask ChatGPT health questions every week, the company says, making it one of the chatbot's most popular use cases.

Users can link several data sources to ChatGPT Health: medical records through partner b.well, Apple Health for exercise and sleep tracking, and wellness apps like Function, MyFitnessPal, and Weight Watchers. AllTrails, Instacart, and Peloton work too.

The feature covers everything from understanding lab results and prepping for doctor visits to nutrition and workout advice. OpenAI makes clear that Health is meant to support medical care, not replace it, and that the tool isn't designed for diagnosis or treatment.

OpenAI promises to keep your medical records out of its training data

Given how sensitive health data is, OpenAI added extra layers of protection. Health runs as a separate area within ChatGPT, with conversations, connected apps, and files stored apart from regular chats. The system even has its own isolated "memories." Context from regular ChatGPT conversations can flow into Health, but Health information never flows back into normal chats. The company also says Health conversations aren't used to train its AI models.

For US health data integration, OpenAI partnered with b.well, a US health data network. Any apps looking to connect with Health must meet OpenAI's privacy and security standards and pass an extra security review.

OpenAI says it spent two years building ChatGPT Health with input from more than 260 doctors who practiced across 60 countries and dozens of specialties. These physicians provided feedback on model outputs in 30 focus areas more than 600,000 times.

Their input shaped how urgently the model recommends seeing a doctor, how clearly it communicates without dumbing things down, and how it handles safety in critical moments. OpenAI measures performance using HealthBench, a custom benchmark that simulates realistic medical conversations. Based on these results, the company claims its latest language models score better than responses from actual doctors.

European users will have to wait

ChatGPT Health isn't available in the European Economic Area, Switzerland, or the United Kingdom yet. OpenAI is starting with a small group of early users and plans to expand access on web and iOS in the coming weeks. For now, only ChatGPT Free, Go, Plus, and Pro subscribers outside Europe can use it.

Some features are also limited by location: medical record integration is US-only, and Apple Health connections require iOS. Europe's stricter data protection rules likely play a role, especially since health data gets extra protections there.

The EU AI Act could also be a factor. A consumer health chat that helps people understand documents and prepare for appointments probably wouldn't fall under the high-risk category of the AI Act. But it could become high-risk if the EU classified it as medical device software.

Whether ChatGPT Health would count as a medical device in the EU depends on how it's marketed and what features it offers. A system that helps with diagnosis faces different requirements than one that explains test results or helps prep for appointments.

The line between information and medical advice is blurry

OpenAI's emphasis that Health isn't meant to replace doctors—just help with preparation and background information—likely reflects legal concerns as much as anything else. But when it comes to health, the line between information and advice gets blurry fast.

This matters especially given how persuasive ChatGPT can be and how easily it could steer users wrong, particularly with mental health issues. For medical questions, that could be dangerous.

Still, there are success stories. In a widely shared Reddit post, a user described how ChatGPT identified a diagnosis that numerous doctors had missed for over ten years. The AI spotted a genetic MTHFR mutation that affects seven to twelve percent of the population.

A study published in the International Journal of Medical Informatics in November 2025 found that ChatGPT-4o outperformed general practitioners on 200 common medical questions, both for accuracy and empathy. The researchers stress that human oversight remains essential to keep AI-generated answers safe and reliable. NYU studies also show that patients often find chatbot responses more empathetic than messages from time-pressed doctors.

The World Health Organization is calling for clear guidelines on transparency, safety, and accountability as AI becomes part of medical care. Some researchers also warn that control over this health data shouldn't rest solely with big tech companies.