According to Fast Company, OpenAI unveiled a new feature called ChatGPT Health on Wednesday. It creates a separate, dedicated space within the chatbot for health questions and data collection. Users can connect wellness apps like Apple Health, Peloton, and MyFitnessPal to it. Crucially, through a partnership with a company called b.well, users can also connect to their electronic medical records, though not directly to systems like Epic’s MyChart. The company says over 230 million people already ask health questions on ChatGPT weekly, and this feature was developed over two years with feedback from more than 260 physicians across 60 countries.
How it works and the big catch
So here’s how they’re trying to make this work. You can funnel your workout data from your Peloton or Apple Watch into ChatGPT Health, giving the AI some context about your activity. For the serious medical stuff—your actual doctor’s notes, lab results, diagnoses—you go through b.well. Basically, b.well acts as a middleman. When you want your records, you’ll likely end up logging into your hospital’s patient portal (which often is MyChart) to authenticate, and b.well fetches that data for ChatGPT.
This is a clever workaround. It means OpenAI doesn’t have to build direct, hospital-by-hospital integrations, which is a nightmare. They let b.well handle the messy, regulated plumbing of health data exchange. But it also adds a step for the user and introduces another company into your sensitive data chain. Is that a trade-off people will accept for the convenience of an AI health assistant? Maybe.
Why this is a major move
Look, the biggest number in that announcement is 230 million. That’s how many people are already using a general-purpose chatbot for health advice every single week. That’s frankly wild and a bit scary. OpenAI is essentially formalizing and trying to corral that behavior into a more controlled, purpose-built environment. By bringing in hundreds of doctors for feedback, they’re trying to bake in guardrails and improve accuracy from the get-go.
And let’s be real: this is a massive market play. Health is the ultimate sticky, valuable, personal dataset. If OpenAI can become the trusted interface for your health information—connecting your fitness tracker data to your lab results to your casual symptom questions—that’s a powerful position. It’s not just about answering questions anymore; it’s about becoming a central hub for your personal health narrative.
The privacy and trust hurdle
Here’s the thing, though. All of this hinges on one massive, unresolved issue: trust. Do people trust OpenAI, a company with a famously chaotic governance structure, with their most sensitive personal data? The b.well partnership is a nod to HIPAA compliance, but it doesn’t automatically solve the perception problem. Every data breach, every internal controversy at OpenAI, will now have a direct line to people’s anxiety about their health privacy.
I think the success of ChatGPT Health won’t be decided by its features first. It’ll be decided by whether OpenAI can convincingly act like a responsible healthcare steward, not just a brilliant tech company. That’s a different culture entirely. They’ve got the doctors on board for the medical part. Now they need to prove they have the maturity for the human part. That might be the harder challenge.
