Two of the major movers in generative and agentic AI announced healthcare products within days of each other–and in time for both the tail end of CES and at JPM. Let’s make some sense out of the hype–along with Claude’s 12-page, ‘connecting the dots’ health plan generated for one individual (see the closing).
First out of the gate was OpenAI entering provider medical management with OpenAI for Healthcare, a set of products that includes ChatGPT for Healthcare. ChatGPT for Healthcare was rolled out earlier this month to AdventHealth, Baylor Scott & White Health, Boston Children’s Hospital, Cedars-Sinai Medical Center, HCA Healthcare, Memorial Sloan Kettering Cancer Center, Stanford Medicine Children’s Health and University of California, San Francisco. ChatGPT for Healthcare is designed to use generative AI to support clinicians in reasoning during patient care and to reduce administrative burden. Key features include:
- Models built for healthcare clinical, operational, and research workflows built on GPT‑5 models
- Evidence retrieval with transparent citations that draw from millions of peer-reviewed research studies, public health guidance, and clinical guidelines
- Institutional policy and care pathway alignment that integrate with enterprise tools such as Microsoft SharePoint and other systems
- Templates to automate workflows
- Access management and governance
- Data control and support for HIPAA compliance
OpenAI’s API has already been used in HIPAA-compliant healthcare software marketed by Abridge, Ambience, and EliseAI. There was no public timetable for availability to other healthcare organizations. OpenAI release
The consumer version, ChatGPT for Health, is still in test. The limited information available indicates that it will provide a secure storage area for connecting personal medical records and wellness apps. ChatGPT is touting its current track record of being a leading source of health and wellness information for 230 million people globally. The new program will help individuals understand recent test results, prepare for doctor appointments, advise on diet and workout routines, and understand the tradeoffs of different insurance options based on healthcare patterns. OpenAI will be operating it as a separate space to protect this data and not use this information to train their models. At this point, it’s waitlisted for nearly everyone, but some users across their various products will be invited to test it. It will be available only in the US. OpenAI release, Healthcare Finance
Days later, Anthropic, the parent company of agentic AI Claude, stole a cheeky march on rival OpenAI at JPM by announcing Claude for Healthcare, not only for providers but also for consumers. The HIPAA-compliant tools for Claude are built on their Opus 4.5 latest software version and on their October release for Life Sciences. The sense this Editor has is that the two actually run somewhat in tandem. Claude’s tools for providers center on prior authorization, insurance claims appeals, care coordination and patient triaging, clinical administration, and healthcare startup developers. Claude’s tools compile information added from the CMS Coverage Database, ICD-10, and the National Provider Identifier (NPI) Registry.
For individuals, Claude Pro and Max subscribers in the US can connect their personal health records and results to Claude. Their integrations or ‘connectors’ include HealthEx and Function in beta, with Apple Health and Android Health Connect integrations following in beta via the Claude iOS and Android apps. Once connected, “Claude can summarize users’ medical history, explain test results in plain language, detect patterns across fitness and health metrics, and prepare questions for appointments. The aim is to make patients’ conversations with doctors more productive, and to help users stay well-informed about their health.” Which puts this connectivity for individuals in the here and now, a step ahead of OpenAI. Anthropic release, Healthcare page
Two 9000-lb elephants in AI have staked out their territory in healthcare. How much takeup both for clinical and personal models will happen, how long they will take to ‘debug’, and how long it will take for a paying clinical model, are interesting bets to take. Anthropic apparently won the PR war by announcing at JPM, as evidenced on CNBC‘s and Mobihealthnews’s reports,. But then there’s this breathless rave review on the Food is Health Revolution blog on how Claude, digesting 60 files collected over a decade, generated a 12-page health plan that connected the writer’s low thyroid with her cognitive problems. Bingo!

















Most Recent Comments