OpenAI rolls out ChatGPT Health in US, warns against diagnostic use

OpenAI launched its ChatGPT Health service on 23 July, making the new AI‑driven health assistant available to all U.S. users aged 18 and over across free, Plus and Pro plans. The offering runs on the GPT‑5.6 Luna model, which OpenAI says outperforms clinicians in certain reasoning tasks, and allows users to link Apple Health data and records from U.S. health systems such as One Medical and Function Health.

During the briefing, Vice President of Health Product Ashley Alexander claimed the models “reason at a level better than a clinician,” a statement quickly tempered by Health Lead Karan Singhal, who emphasized that the service is not intended for diagnosis or treatment. OpenAI’s terms of service explicitly prohibit using ChatGPT for medical diagnosis or treatment, a clause the company has cited in its defense of a lawsuit filed the day before the launch.

Former Florida pastor Scott Winters sued OpenAI and CEO Sam Altman in San Francisco Superior Court, alleging that a July 2025 interaction with ChatGPT‑4o caused him to postpone a doctor’s visit, leading to a massive pulmonary embolism. The case highlights the tension between the product’s promotional claims and its legal restrictions.