Anthropic Adds Features For Doctors, Patients In Health Care Push

Anthropic is making it easier for patients and clinicians to use its artificial intelligence chatbot to access medical information, part of a broader push into the lucrative health care sector.

The San Francisco-based company on Sunday said that its Claude product is launching a new health care offering that is compliant with the Health Insurance Portability and Accountability Act, or HIPAA, and can be used by hospitals, medical providers and consumers to field protected health data. Anthropic has also integrated scientific databases into its product and added enhanced capabilities for biological research.

On the consumer front, Anthropic is allowing users to export their health data from apps including Apple Health and Function Health, with the goal of helping them gather and share medical records with providers.

Anthropic, which is currently in funding talks at a $350 billion valuation, announced the features days after rival OpenAI unveiled new tools for clinicians to work through cases and for everyday users to review their test results, diets and workout routines. The back-to-back releases highlight Silicon Valley’s growing desire to gain ground in health care to bolster sales and prove AI’s broad benefits.

“When we think about the overall economy and where AI can have the most impact, it is really well primed for that, once you do the right things on the regulatory and data front,” said Mike Krieger, Anthropic’s chief product officer and a co-founder of Instagram. The new tools are designed for “empowering people to have more knowledge, both from their data, but also in conversation with their providers,” Krieger said.

Founded in 2021 by former employees of OpenAI, Anthropic has positioned itself as a reliable, safety-conscious AI developer. Its software has become particularly popular among engineers who use it to automate the coding process. But Anthropic, whose chief executive officer, Dario Amodei, is a biophysicist by training, has also started to see some early traction from medical providers.

The company said Banner Health, one of the largest nonprofit health systems in the US, has more than 22,000 clinical providers using Claude, and that 85% report working faster with higher accuracy. Anthropic is also working with customers such as Novo Nordisk A/S and Stanford Health Care.

The Claude maker faces steep competition not just from OpenAI but also legacy technology providers as well as newer startups, which have tried to apply advances in AI to drug discovery, medical paperwork and analyzing patient records. But these ventures also come with new privacy and safety risks from AI handling sensitive personal data and offering suggestions for high-stakes health matters.

Anthropic said its medical responses are grounded with citations from respected publications such as PubMed and the NPI Registry, ensuring that clinicians can have more confidence in the results. Anthropic also said it will not train its models on health care user data.