Karan Singhal Drives ChatGPT Health Advice Improvements

Business Insider reports that Karan Singhal, who leads health AI at OpenAI, is driving the company's push to improve ChatGPT's health and wellness advice. Business Insider cites OpenAI's own figure that more than 230 million people ask ChatGPT health and wellness questions every week, and reports Singhal's claim that OpenAI's most recent model was the first trained "at every stage of development" to better handle health questions. Singhal joined OpenAI in mid-2024 after working on medical models at Google. Business Insider also reports that GPT-4o has faced lawsuits alleging it gave harmful health advice, which OpenAI has denied.

The push follows OpenAI's January 7, 2026 launch of ChatGPT Health, a dedicated in-app experience that lets users connect medical records and wearable data from services like Apple Health and MyFitnessPal, according to Fortune's reporting from that launch. At the time, Singhal said OpenAI had been laying the groundwork for the product for about two years, and OpenAI's applications CEO Fidji Simo said the effort sat outside the company's separate "code red" initiative. OpenAI has said it does not train its models on users' personal medical data and does not describe ChatGPT Health as HIPAA-compliant, since consumer health apps generally fall outside HIPAA's scope.

Teams building on general-purpose LLMs for health use cases sit in the same regulatory gray zone OpenAI does: HIPAA generally does not apply to consumer-facing apps like ChatGPT, which puts the compliance burden on layered internal safety controls rather than a pre-existing legal framework. OpenAI's approach, partnering with a health-data platform (b.well) rather than building EHR integrations from scratch, and keeping health conversations out of model training by default, are patterns worth watching as more vendors move into this space; Google made a similar b.well partnership in October 2025 without yet shipping a dedicated Gemini health product.

Whether OpenAI publishes technical benchmarks or third-party clinical evaluations for the newest model's health-specific accuracy and safety, since neither Business Insider nor Fortune report that it has; and how the GPT-4o lawsuits Business Insider references develop, since the outcomes could shape how aggressively OpenAI and competitors market health features going forward.