OpenAI Opens ChatGPT Health to All U.S. Adults as Scrutiny Over Medical Advice Grows

OpenAI is expanding ChatGPT Health to all logged-in U.S. users age 18 and older, making its health-focused features available across free, Go, Plus and Pro plans on the web and iOS. The rollout matters because it brings one of the company’s most sensitive AI features into the hands of a much wider audience just as regulators, doctors and critics are questioning whether chatbots are safe enough to guide medical decisions.

The move comes only a day after a Florida pastor filed suit against OpenAI, alleging the chatbot gave dangerously bad medical guidance and discouraged a doctor visit. OpenAI says it is trying to make health responses safer, more useful and easier to personalize, but it continues to warn users not to rely on the system for diagnosis or treatment.

What OpenAI launched and who can use it

OpenAI said ChatGPT Health is now available to all U.S.-based adults who are logged in, rather than being limited to a smaller test group. The feature is rolling out this week on the web and in the iOS app, and it is included for users on free, Go, Plus and Pro plans.

The company first introduced the health hub in January as a dedicated place for people to ask health-related questions and connect outside data sources. At the time, OpenAI positioned the feature as a way to make ChatGPT more practical for everyday wellness questions while keeping more sensitive requests in a more controlled environment.

How the feature works

ChatGPT Health can draw on information users connect from consumer health apps, personal tracking tools and medical record systems. OpenAI says supported sources include Apple Health, Function and MyFitnessPal, along with provider systems such as Epic and Oracle Health and services like One Medical and Function Health.

That means a user can ask a broader question in the general chat and still have ChatGPT combine those connected details into its response. In practical terms, someone could ask about a food choice, a symptom pattern or an allergy concern and have the model consider health data they have already shared with the service.

Why OpenAI changed the experience now

OpenAI says the new rollout reflects how people are already using ChatGPT. During testing, the company observed that 70% of health-related prompts were being asked outside the dedicated health hub, suggesting many users wanted answers in the main chat interface rather than in a separate section.

That usage pattern appears to have helped push OpenAI toward a broader integration strategy. Rather than forcing users into a special destination for health questions, the company is now allowing connected health data to inform replies anywhere in ChatGPT’s general experience.

What changed from the January test

Earlier in the year, OpenAI had positioned the health hub as the primary place to ask questions about medicine, fitness and wellness. The company had also reported that users were submitting 230 million health-related queries each week.

OpenAI now says that figure has climbed to 300 million weekly health-related queries, underscoring the scale of demand. The larger number helps explain why the company is moving quickly to give more users access, even while keeping some guardrails in place.

• January: OpenAI begins testing ChatGPT Health through a dedicated hub.
• During testing: 70% of health prompts appear outside the hub.
• At launch: 230 million weekly health queries are reported.
• Now: OpenAI says the figure is 300 million per week.
• July 2026: The feature expands to all logged-in U.S. adults over 18.

How much trust should users place in AI health answers?

Users should treat ChatGPT Health as a starting point, not a medical authority. OpenAI itself says its service is not intended to diagnose or treat any condition, and it is urging people to verify information and consult trained clinicians before making medical choices.

The tension is obvious: the company is promoting a more personalized health experience while simultaneously warning that the output should not be used as medical advice. That contradiction is at the center of the broader debate over AI in healthcare, where convenience and access often clash with accuracy and liability concerns.

OpenAI says it is working with physicians to improve health-related responses, and it continues to advise users to confirm important information with medical professionals rather than treating the chatbot as a substitute for care.

OpenAI also says it does not use user data to train its models, a point likely intended to reassure people who are wary of sharing private medical information with a consumer AI product. Still, the company’s own rollout language makes clear that any health answer should be treated as informational rather than clinical.

What does OpenAI say about model performance?

OpenAI argues that its latest models are better at health-related tasks than earlier versions. The company says GPT-5.6-Luna, described as the smallest model in its most recent release family, performs better than GPT-5.5 on HealthBench, an open-source benchmark OpenAI created to evaluate large language models on health queries.

Benchmarks like HealthBench matter because they are among the few public tools for comparing how models handle sensitive domains. But performance on an evaluation set does not automatically translate into safe or reliable behavior in the real world, especially when users ask open-ended questions about symptoms, medications or emergency situations.

Why benchmarks matter in healthcare AI

Health-related AI systems are judged on more than fluency. They have to balance factual accuracy, uncertainty, caution and context. A model can sound confident while still being wrong, which is one reason healthcare is such a difficult test for general-purpose AI systems.

OpenAI’s emphasis on HealthBench is also a signal to the market. The company is trying to show that it is not simply widening access to a consumer feature, but grounding the rollout in measurable progress. That is a persuasive message for investors and users, though it may not be enough to satisfy critics worried about real-world harm.

What happened in the lawsuit and why does it matter?

The timing of the release is notable because it follows a legal complaint filed by a Florida pastor who says ChatGPT gave a near-fatal recommendation and discouraged him from seeking a doctor. The case adds pressure to OpenAI at exactly the moment it is opening a more capable health feature to a much larger audience.

Although the company has not said the launch was prompted by the lawsuit, the overlap is hard to ignore. OpenAI has already pointed to its service terms in legal responses, emphasizing that the product is not a medical device and is not meant to replace professional care.

The dispute highlights a broader problem in consumer AI: as models become more useful for everyday questions, users may become more likely to trust them in situations that carry real health risk. That is especially true when a system appears to blend personal data, conversational convenience and confident language into one interface.

How do competitors approach health AI?

OpenAI is not alone in pushing into health-related AI features. Rival companies including Anthropic and Google have also introduced or explored health-oriented tools, reflecting a wider industry belief that medical questions are one of the biggest consumer uses for generative AI.

The appeal is straightforward. People already use chatbots to interpret lab results, compare symptoms, decode medication instructions and prepare for doctor visits. Tech companies see an opportunity to make those interactions faster and more personalized, even as public-health experts warn that the results can be inconsistent or misleading.

Why the category is so contentious

Health sits at the intersection of user demand, regulatory uncertainty and high-stakes consequences. A wrong answer about dinner choices may be harmless. A wrong answer about chest pain, medication interaction or a child’s symptoms can be dangerous.

That is why many studies have warned against depending on AI bots for medical advice. The systems can miss nuance, fail to recognize urgency or offer generic guidance that does not fit a specific person’s condition. Yet the consumer appetite for quick answers keeps growing, and that demand continues to pull companies deeper into the space.

What personal data can ChatGPT Health use?

ChatGPT Health can incorporate user-connected health information from multiple sources if the user authorizes it. The feature is designed to work with consumer apps, health platforms and some medical record systems, which makes it more flexible than a simple question-and-answer chatbot.

Among the sources OpenAI names are:

• Apple Health
• MyFitnessPal
• Function
• Epic
• Oracle Health
• One Medical
• Function Health

The idea is to make ChatGPT more aware of an individual’s context, such as diet patterns, allergies, exercise routines or clinical records that the user chooses to connect. That may make responses more relevant, but it also raises obvious privacy questions around sensitive health information, access control and user understanding of what is being shared.

What this means for privacy and security

Any feature that blends conversational AI with medical records demands a high level of trust. Even when a company says it does not train on user data, the act of connecting health accounts can still create anxiety about storage, retention, access permissions and misuse.

OpenAI has not presented the rollout as a replacement for healthcare providers, but the feature’s design encourages users to bring more personal context into a consumer product. That may improve relevance, yet it also increases the potential consequences if something goes wrong.

For users, the practical question is whether the convenience is worth the risk. For OpenAI, the challenge is proving that the company can safely scale a health product without normalizing overreliance on AI for diagnosis, treatment or urgent care decisions.

Why this rollout is strategically important for OpenAI

The health feature is part of a larger push to make ChatGPT more embedded in daily life. Health is one of the few domains where consumers routinely seek highly personal, repeated guidance, which gives the company a strong reason to improve the product’s utility and retention.

At the same time, expanding access broadens the company’s exposure. The more people use ChatGPT for medical concerns, the more OpenAI has to answer questions about reliability, data handling and legal responsibility. That combination makes health one of the most commercially attractive and reputationally risky areas in consumer AI.

Where the product could go next

OpenAI has not announced a roadmap beyond the U.S. adult rollout, but the current direction suggests a few likely priorities. Those may include deeper integrations with additional health providers, better contextual reasoning around connected records and stronger safeguards around emergency language and high-risk topics.

It is also likely that OpenAI, like its competitors, will continue to invest in evaluations and physician input to show that its models are improving. The company’s challenge will be making those gains visible to users without encouraging them to assume the system is clinically authoritative.

Bottom line

OpenAI is widening access to ChatGPT Health at a moment when demand for AI medical guidance is surging and concerns about safety are intensifying. By opening the feature to all logged-in U.S. adults over 18, the company is betting that more personalized answers will outweigh the risks of using a chatbot for sensitive health questions.

That bet will now be tested at scale. Users will likely embrace the convenience, clinicians will continue warning against overtrust, and OpenAI will have to prove it can make the product more useful without making it more dangerous.