The U.S. poured $4.9 trillion into health care in 2023, equal to roughly 17.6% of GDP, federal data show.

What jumps out from that number is not just the size, but how much of it is tied up in confusion, bureaucracy, and fragmentation that ordinary people feel every time they try to book an appointment or read a bill, according to an OpenAI analysis of patient pain points.

ChatGPT Health is essentially a bid to sit at the front door of that system by becoming the place where you upload your lab results, connect your Apple Health data, and ask what that new diagnosis or bill really means.

OpenAI describes it as a “dedicated experience that securely brings your health information and ChatGPT’s intelligence together, to help you feel more informed, prepared, and confident navigating your health,” according to the company’s launch announcement.

What ChatGPT Health actually does

ChatGPT Health lives in its own space inside ChatGPT, with separate memories, encryption, and isolation, so health conversations are walled off from everything else you ask the chatbot, OpenAI said.

“Health lives in its own space within ChatGPT, where your conversations, connected apps, and files are stored separately from your other chats,” the company said, adding that those chats are not used to train its foundation models.

Once users are onboarded, they can link electronic medical records through U.S. data partner b.well, as well as apps like Apple Health, MyFitnessPal, Function, Weight Watchers, and Peloton.

The company says that means people can ask questions like “How’s my cholesterol trending?” or “Can you summarize my latest bloodwork before my appointment?” and get answers grounded in their actual history rather than generic web advice.

Why OpenAI is moving toward health now

Health has already become one of ChatGPT’s biggest organic use cases, with OpenAI estimating that more than 5% of all global messages are about health care and that more than 40 million people turn to the tool daily with health questions.

The shift will force health systems to catch up with patient‑facing AI standards, argued Vital CEO Aaron Patzer in an interview with Chief Healthcare Executive. 

At the same time, the U.S. health care system is both expensive and deeply unpopular, with three in five Americans viewing it as “broken” and hospital costs as the top complaint, according to polling cited in an OpenAI‑linked report.

That combination of entrenched frustration and rising comfort with consumer AI is exactly the kind of opening tech companies look for when they say they want to reshape an industry, especially one that already accounts for $4.9 trillion in annual spending, according to federal figures.

Patients are already using ChatGpt for their health records.

How ChatGPT Health tries to address trust and safety

OpenAI is careful to say that ChatGPT Health is “designed to support, not replace, medical care” and is “not intended for diagnosis or treatment.”

Instead, the tool is positioned as a way to explain lab results in plain language, prepare questions for doctors, interpret wearable data, and navigate insurance tradeoffs, particularly in confusing areas like plan selection and surprise bills, OpenAI said.

Some founders in the space say the scope of that ambition is hard to overstate.

Rainmatter Capital Investment Lead Dilip Kumar shared his thoughts on ChatGPT Health in a post on X (formerly Twitter).

To build credibility, OpenAI spent two years working with more than 260 physicians across 60 countries, collecting more than 600,000 pieces of feedback on model outputs across 30 clinical focus areas.

That physician‑led evaluation process fed into “HealthBench,” an internal rubric that scores responses on safety, clarity, and how well the model escalates people to real‑world care when it detects potential red flags. 

Why health care incumbents should care

For an industry used to slow, regulation‑heavy change, user behavior around generative AI is moving fast.

Health care has become one of the fastest‑growing sectors for ChatGPT Enterprise usage, with an eight‑fold increase as hospitals and health systems embed tools into back‑office and clinical workflows, according to OpenAI’s enterprise report, cited by Mobi Health News. 

More Health Care:

  • If your Medicare plan was canceled, do this now
  • Health care costs are the wild card in year-end tax planning
  • 22 million Americans hit by ACA health insurance cliff after vote fails
  • How Work Conditions Impact Mental Health

“In 2026 we will see AI shift from isolated pilot programs to full enterprise-scale deployment, driven by clearer return on investment,” Ben Sharfe, EVP for AI at Altera Digital Health, told Chief Healthcare Executive, pointing to ambient listening and native EHR integrations as likely tipping points.

Consumer behavior is shifting just as quickly on the front end. More than 70% of health‑related ChatGPT conversations happen outside clinic hours, and hundreds of thousands of messages each week come from U.S. “hospital deserts” more than 30 minutes from a hospital, according to OpenAI’s analysis.

That pattern suggests people are already treating chatbots as a kind of after‑hours triage and navigation layer, even before a purpose‑built product like ChatGPT Health fully rolls out.

Health and AI: where the money is

If you strip away the hype, what ChatGPT Health really targets are the messiest, least-loved aspects of the $4.9 trillion system: administrative friction, opaque billing, and information overload, health economists note.

Researchers have long pointed out that a substantial share of U.S. health spending goes to administration and overhead rather than direct clinical care, a gap that technology vendors say AI can narrow by automating documentation, coding, scheduling, and claims workflows.

Related: Novo and Lilly shift GLP-1 strategy abroad: US may be next

Here’s where generative AI is already being deployed in health care:

  • Automating clinical documentation and medical coding to cut billing and insurance‑related overhead.
  • Streamlining scheduling, intake, and claims adjudication to reduce delays and denials.
  • Surfacing understandable explanations of records and benefits to reduce information overload for patients.​

In parallel, consumers are now spending heavily on wellness, with the U.S. wellness market alone estimated at around $2 trillion and representing roughly one‑third of the global wellness economy, according to the Global Wellness Institute. 

When I look at how ChatGPT Health integrates Apple Health, workout apps, and nutrition trackers alongside clinical records, it feels like a deliberate attempt to sit at the intersection of that wellness economy and the formal health care system, rather than choosing one side or the other.

What ChatGPT Health could mean for patients and investors

For patients, the upside is obvious: a single place to see what your labs mean, whether that bill makes sense, and what questions to bring to your doctor tomorrow, instead of juggling portals and PDFs, OpenAI argues.

“Americans are using AI and ChatGPT to equip themselves with information to gain more agency over their health,” arguing that better information can help people challenge denials, spot billing errors, and make more informed coverage choices.

For investors and incumbents, the bigger story is whether a consumer‑facing chatbot can become a de facto interface for pieces of a $4.9 trillion system that has resisted transformation for decades.

If ChatGPT Health succeeds in becoming the default way millions of people interpret records, pick plans, and prepare for visits, it could quietly shift power away from traditional gatekeepers toward whoever controls that interface, even if the underlying providers, insurers, and regulators remain the same.

Related: Michael Burry shares bold predictions for OpenAI, Palantir