For years, when something felt off, the first thing you did was open Google and type in your symptoms. A headache became a brain tumour. A rash became something far worse.
Now, instead of scrolling through blue links and panic-inducing forums, people are turning to ChatGPT for answers.
While ChatGPT has become a go-to for health queries, it was never designed to handle sensitive medical questions. It hallucinates, is biased based on your questions and may give wrong advice.
In response to those limits, last week, OpenAI launched ChatGPT Health. A direct push into healthcare with a promise of clearer explanations, better continuity of care, and stronger privacy safeguards.
The goal is to help people understand their health before, during, and after their doctor consultations.
ChatGPT Health: A digital consultation room
ChatGPT Health is not a separate app. It is a dedicated tab inside the ChatGPT interface, designed to feel like a private digital consultation room. Here, users can upload lab reports, review medical documents, and get plain-language explanations of what those numbers actually mean.
For U.S. users, ChatGPT Health connects with health data platforms like b.well, which securely links medical records from hospitals and providers. It can also sync with wellness and fitness apps such as Apple Health, MyFitnessPal, Peloton, and Oura, pulling fragmented data into a single view.
The core features include:
Lab report analysis
When you upload a lab report, it translates your complex medical data and helps you understand the biomarkers, their ranges, and what the report tells about your health.
Appointment preparation
ChatGPT can review your recent health queries and suggest focused, high-value questions to ask your doctor.
Holistic health tracking
By combining sleep, activity, nutrition, and vitals from wearables and apps, it can surface correlations like how your poor sleep spiked your heart rate.
Insurance navigation
It helps decode dense insurance documents and benefits language.
As cardiologist and digital health advocate Dr. Eric Topol puts it: “AI can restore the human element of medicine by handling the data crunching.”
OpenAI’s aim
According to Fidji Simo, OpenAI’s VP of Product, ChatGPT Health is designed to address one of healthcare’s biggest structural problems: continuity of care.
In today’s system, your dermatologist, primary care doctor, and nutritionist often operate in silos, with little visibility into each other’s decisions. OpenAI wants ChatGPT Health to act as a connective layer where patients can see the full picture.
“We aren’t replacing doctors,” Simo said at launch. “We are empowering patients to be the CEO of their own health.”
By providing patients with better information, OpenAI believes it can reduce administrative burnout for doctors and improve outcomes for patients who feel lost inside complex healthcare systems.
That vision is already resonating with institutions. John Brownstein, SVP and Chief Innovation Officer, Boston Children’s Hospital, said
“Our early work with a custom OpenAI-powered solution allowed us to move quickly, prove value in a secure environment, and establish strong governance foundations.
ChatGPT for Healthcare offers a path toward operational scale, providing an enterprise-grade platform that can support broad, responsible adoption across clinical, research, and administrative teams.”
Concerns with ChatGPT Health
Security
Health data is among the most sensitive info anyone can share. While any AI product has security risks, OpenAI promises not to train its future models on data shared with the Health tab.
The company says ChatGPT Health will operate differently from regular ChatGPT. Health-related chat history will stay isolated within the user’s account and will not be collected by AI.
The feature uses purpose-built encryption and isolation, ensuring that when you switch to the ChatGPT health tab, you enter a different security container.
For integration with external apps and providers, OpenAI relies on b.well, a HIPPA-Compliant platform that acts as a secure pipe between hospitals, apps, and ChatGPT Health.
Privacy
Connecting hospital records, nutrition apps, and fitness trackers in one place creates an extremely valuable and sensitive data profile.
Users also need to pay close attention to third-party app policies. While OpenAI may not train models on health data, ChatGPT can still send summaries to connected apps, potentially expanding the surface area for data exposure.
Even OpenAI CEO admits that AI platforms cannot offer doctor-patient confidentiality legally, which is a key limitation.
AI hallucinations
While OpenAI has added guardrails and inputs from more than 260 physicians, ChatGPT is still a language model.
It predicts likely responses and does not “know” medicine. As a result, it can still make things up.
To address this, OpenAI has introduced tighter controls in the health tab:
- Clear disclaimers reminding users that the model can make mistakes: “ChatGPT can still make mistakes. Check important info with a doctor.”
- Refusals to provide diagnoses or treatments in high-risk situations, instead directing users to emergency services.
- More aggressive citation of medical journals and authoritative sources, with less reliance on general web content.
Still, hallucinations remain a risk and potentially a dangerous one in healthcare.
What this means for users
At its best, ChatGPT Health reduces friction. It can help you:
- Understand lab reports without hours of research
- Remind you about medications
- Help plan diets around allergies and cholesterol levels
But at its worst, it could encourage over-reliance on numbers, metrics, and machine-generated reassurance.
Cautions users need to keep in mind
- Emergency situations: If you are bleeding, having trouble breathing or in severe pain, ChatGPT Health is useless. You must reach out for medical help.
- Data minimisation: Connect only the apps and records that you need.
- Verification: Always confirm ChatGPT suggestions from licensed physicians and pharmacists before acting on them.
What about developers
ChatGPT Health is currently closed to third-party developers. There is no open “health app store,” and that appears to be a deliberate safety choice. Any plugin operating in this space would require scrutiny closer to FDA-level oversight.
That said, future APIs are likely. OpenAI may eventually allow compliant health apps to feed data into ChatGPT Health, provided they meet strict security and governance standards.
The bottom line
ChatGPT is already changing how people engage with healthcare information. OpenAI’s zero-training privacy promise is a meaningful step, but trust in healthcare is fragile, earned slowly and lost instantly.
If OpenAI avoids major hallucination failures or data leaks, ChatGPT Health could become a powerful bridge between patients, health data, and doctors. If not, it will serve as a reminder that when it comes to health, there is no substitute for human expertise.
-By Dr Rohini Devi and the AHT Team
