According to Fortune, OpenAI has announced ChatGPT Health, a new dedicated experience where users can connect medical records and apps like Apple Health and MyFitnessPal. The company’s CEO of applications, Fidji Simo, introduced the feature by sharing a personal story where ChatGPT flagged a dangerous drug interaction for her last year. OpenAI has been laying groundwork for this for about two years, led by high-profile healthcare hires like Nate Gross from Doximity. They found that over 230 million people globally ask health-related questions on ChatGPT every week. The feature is launching via a waitlist, with a full rollout on web and iOS in the coming weeks, though EHR integrations are U.S.-only for now. OpenAI also clarified this project was not part of its recent internal “code red” sprint.
The Personal Data Gamble
Here’s the thing: this is a massive trust exercise. OpenAI is explicitly saying it won’t train its models on your personal medical data, which is the absolute bare minimum. But they’re asking users to pipe in their most sensitive information—medical records, financial data from platforms like partner b.well, fitness tracking—into a chat interface. They’re adding layers like multi-factor authentication and letting you revoke access, which is good. But let’s be real, the real test is whether people feel comfortable doing this after the company’s historically aggressive approach to training data. It’s a necessary pivot, but the skepticism is baked in.
A Crowded Race With No Clear Leader
OpenAI is not pioneering here; they’re joining a stampede. Google just partnered with the same company, b.well, last October. Apple has decades of health data locked in its ecosystem. Amazon has tried this. Countless startups are at it. So what’s OpenAI’s edge? Basically, it’s the existing user base and the conversational interface. 230 million weekly health queries is a staggering number, and it shows people are already using it as a WebMD on steroids. The bet is that by making those conversations smarter with your actual data, they become indispensable. But it’s a fast-moving race, and being first to chat doesn’t mean you’ll win the whole healthcare marathon.
The Vertical Playbook
This isn’t just about health. Look at the pattern. Study Mode for education. Agentic shopping features. A rumored finance product. OpenAI is methodically building vertical-specific experiences on top of its general-purpose model. The era of “just talk to the AI about anything” is maturing into “talk to the AI *for* specific things.” It’s a smart product strategy to increase engagement and utility, moving from a cool toy to a professional tool. Health is just the most sensitive and, potentially, most impactful vertical they’ve tackled yet. If they can get this right, the blueprint for other industries becomes much clearer.
The Real-World Test
Simo’s kidney stone story is a powerful anecdote, but it’s also a perfect example of the promise and the peril. The AI connected dots a time-pressed human missed. That’s the dream. But the nightmare is a hallucination or a misinterpretation with the same high stakes. The press release phase is easy. The real launch happens when millions of non-CEOs start relying on it for advice about medications, symptoms, and chronic conditions. The regulatory landscape is fuzzy, the competition is fierce, and the margin for error is effectively zero. OpenAI has built the feature. Now we see if the world decides to use it.
