OpenAI Launches ChatGPT Health Amid Safety Lawsuits
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI rolled out "Health in ChatGPT" this week, letting users upload medical records like lab results and medications and connect Apple Health for fitness and sleep data
- OpenAI said 300 million people already turn to ChatGPT with health questions weekly — a volume Cornell Tech's Tanzeem Choudhury attributes to "broken" U.S. healthcare access barriers
- A Florida man sued OpenAI this week alleging ChatGPT's medical advice nearly killed him during a pulmonary embolism, months after a California teenager died after ChatGPT said it was safe to consume a fatal concoction of substances
- OpenAI said physicians helped develop realistic health scenarios and test the product, adding encryption safeguards, user check-ins before connected plugins act, and pledging uploaded records won't train models or target ads
- Cornell Tech's Tanzeem Choudhury noted AI companies have quietly changed terms of service before and pointed to an FTC warning that such changes can be "unfair or deceptive" to consumers
- Columbia University's Vishal Misra said he personally uses ChatGPT to track his own health data but urged users to always take AI advice to a real physician before acting
Why it matters: With 300 million weekly health queries already flowing through ChatGPT, OpenAI is formalizing a medical-adjacent product while two recent harm cases — a near-death lawsuit and a teen's death — sit in the public record. Choudhury's warning that AI firms have quietly rewritten ToS before puts OpenAI's "records won't train models" pledge on a short clock the moment the policy shifts.
