Doximity Outranks OpenEvidence, Frontier Models in Independent Stanford-Harvard Study of Clinical AI Safety

Doximity Outranks OpenEvidence, Frontier Models in Independent Stanford-Harvard Study of Clinical AI Safety reports that Doximity’s HIPAA-compliant Doximity Ask outperformed leading frontier AI systems on the NOHARM benchmark, in an independent evaluation. The study was conducted by ARISE, a clinical AI research team led by physicians from Stanford and Harvard Medical Schools, using simulated patient cases. Doximity Ask ranked first on the benchmark’s real-world clinical sample, the segment most similar to physician usage, while purpose-built clinical AI systems generally beat general-purpose frontier models. Doximity attributes its performance to physician authorship, citing its PeerCheck program, where more than 11,000 cited physician experts evaluated and improved outputs. The company also says its Clinical AI Suite has been deployed and approved across more than 150 health systems, including eight of the nation’s top 20 hospitals, with security features such as encryption, role-based access controls, audit logging, and session isolation.