OpenAI Rolls Out ChatGPT Health With Clinician-Level Claims and a Pending Lawsuit
OpenAI has expanded ChatGPT Health to all US users, letting them query personal medical records inside the main app, even as a lawsuit alleges the tool gave one user dangerously wrong advice.
OpenAI has expanded ChatGPT Health across the United States, letting users connect electronic medical records and fitness tracker data directly inside the main ChatGPT interface. As The Verge reports, individuals can now query post-visit notes, lab results, prescribed medications, and personal health history without switching to a separate application tab.
At a press briefing for the rollout, Ashley Alexander, OpenAI's vice president of health product, stated that the company's models reason at levels better than human clinicians. Karan Singhal, OpenAI's health lead, later qualified that claim, saying the superior-reasoning framing requires nuance, while pointing to recent research from Harvard University and Stanford University as evidence that model capabilities are aligning with clinical standards.
The tool runs on GPT-5.6 Sol, a specialized model OpenAI launched specifically for medical applications. The company says user consent is required before any personal record data is analyzed, and that all medical conversations and integrated health records are protected by enhanced encryption in transit and at rest.
The broad release lands alongside active legal scrutiny. A Florida pastor has filed a lawsuit against OpenAI, alleging that ChatGPT gave him dangerously inaccurate medical advice that led him to delay treatment for a pulmonary embolism. The case has sharpened concerns among medical professionals about hallucinated diagnostic guidance and the risks of patient reliance on unverified software outputs.
The Cognarah Angle
The appeal of this product for Africa is obvious. Nigeria has fewer than four licensed physicians for every 10,000 citizens, and primary healthcare infrastructure in rural areas remains chronically underfunded across sub-Saharan Africa. A tool that puts clinical-grade reasoning into the hands of ordinary citizens sounds, at first pass, like a meaningful shortcut around systemic failure.
It is not that simple. AI health models trained predominantly on Global North medical data carry real diagnostic risk when applied to African populations. Disease burdens across the continent, including high prevalence of tropical infectious diseases, distinct genetic profiles, and localized drug supply realities, are rarely well-represented in foreign training datasets. A model that performs adequately on conditions common in American or European clinical records may perform poorly on presentations that are routine in Lagos or Nairobi.
The liability question is equally unresolved. If a user in a well-resourced Western healthcare system suffers serious harm because a chatbot hallucinated a diagnosis, the consequences in a resource-constrained environment could be categorically worse. Regulators including Nigeria's National Information Technology Development Agency and Kenya's Data Protection Commissioner need to move first, establishing explicit standards for automated healthcare platforms before these tools find their way into local markets without validation.
OpenAI has not demonstrated that ChatGPT Health is safe for users in its home market. Releasing it into health systems that carry higher stakes, with fewer fallback options, is not a step toward healthcare equity. It is an uncontrolled experiment with no local ethics board in sight.
Reporting sourced from The Verge. Analysis and Cognarah Angle are Cognarah's own.
Written by
StacyAI-assisted news curation. Every story is reviewed by our editors before publication.
