OpenAI Expands ChatGPT Health to All US Users, Claims Clinical-Level Reasoning

OpenAI is expanding access to ChatGPT Health across the US market, integrating medical record analysis into the main chat interface and positioning the service as a clinical-grade reasoning tool—even as a lawsuit alleges the system dispensed dangerous health guidance that delayed critical care.

ChatGPT Health Goes Nationwide on July 24

According to The Verge AI, OpenAI rolled out ChatGPT Health to all logged-in US users aged 18 and older on Thursday, July 24, 2026, across web and iOS platforms. The feature is available at all subscription tiers—free, Go, Plus, and Pro plans. Previously limited to a dedicated tab, health queries now route through the main chat experience, allowing users to ask contextual questions about connected medical information. Users can integrate records from healthcare providers, Apple Health, MyFitnessPal, Weight Watchers, and Function, with granular controls to review and adjust which data sources feed into ChatGPT.

The “Clinician-Level Reasoning” Claim and Its Qualification

During the launch briefing, OpenAI Vice President of Health Product Ashley Alexander asserted that the company’s models “are now capable of reasoning at levels that are better than clinician level.” The Verge reports that when pressed for evidence, OpenAI Health Lead Karan Singhal walked back the claim, saying he would “temper” the assertion while acknowledging “individual studies…pointing in that direction,” citing research from Harvard and Stanford. This hedging—between a bold opening claim and a more cautious restatement—mirrors a pattern in AI health announcements where marketing assertions encounter scrutiny during technical Q&A.

Data Security and Liability Disclaimers

OpenAI states that all ChatGPT Health conversations use end-to-end encryption, with additional protection layers for connected medical information. Users grant explicit permission before health data personalize responses. The company’s terms emphasize the service “supports, not replaces, professional care”—a standard liability disclaimer that may offer limited protection given the legal exposure emerging in real time.

Concurrent Litigation and Safety Questions

On Wednesday, one day before the nationwide rollout, a Florida pastor filed suit against OpenAI alleging that ChatGPT provided “extremely dangerous medical recommendations” that caused him to delay treatment for a pulmonary embolism, a potentially life-threatening blood clot. According to The Verge, the timing raises questions about the robustness of safety guardrails in a system being marketed for clinical-adjacent reasoning.

Why This Matters

OpenAI’s health expansion tests the boundary between AI assistance and medical advice at scale. If the Harvard-Stanford studies hold up under independent reproduction, the service could reduce friction for patients seeking preliminary guidance on their conditions. However, the concurrent lawsuit signals that disclaimers and encryption may not shield the company from liability when users rely on ChatGPT for time-critical decisions. Hospitals, insurance companies, and regulators will likely scrutinize whether “supports, not replaces” language is sufficient when the interface design—integrating health data into a conversational flow—invites users to treat ChatGPT as a diagnostic tool. The outcome will shape whether AI health features expand into mainstream care workflows or remain confined to non-critical advisory roles.