Doximity Ask tops frontier AI models in NOHARM safety benchmark
Doximity, Inc. announced that Doximity Ask outperformed leading frontier AI models in the NOHARM benchmark, an independent evaluation of clinical AI safety conducted by ARISE. The platform ranked first on a real-world clinical sample, with purpose-built systems outperforming general-purpose models across 1,100 scenarios. Doximity attributes this success to its clinical workflow design and the PeerCheck program involving over 11,000 physician experts.

*this image is generated using AI for illustrative purposes only.
Doximity, Inc. announced that its HIPAA-compliant clinical AI platform, Doximity Ask, outperformed leading frontier AI models in the NOHARM benchmark. This independent evaluation of clinical AI safety was conducted by ARISE, a clinical AI research team led by physicians from Stanford and Harvard Medical Schools. The study assessed how AI models handle simulated patient cases to determine their safety and reliability in medical settings.
Doximity Ask ranked first among all AI systems evaluated on the study's real-world clinical sample, which mirrors how physicians use these tools in practice. Purpose-built clinical AI systems outperformed general-purpose frontier models across the broader automated evaluation. The benchmark included automated testing across more than 1,100 scenarios spanning 10 medical specialties and was developed by over 50 researchers with contributions from 29 board-certified physicians.
Performance Drivers
Doximity attributes its performance to the platform's design, which is built specifically for clinical workflows. The company emphasized that its results stem from investment in physician authorship at scale. Through its PeerCheck program, more than 11,000 cited physician experts have evaluated and improved Doximity Ask outputs.
"We have long believed that the path to trustworthy healthcare AI runs through physicians, not around them," said Dr. Louis-Antoine Mullie, Head of Medical AI at Doximity. "Continuous physician review isn't a differentiator. It's a requirement. This result reinforces the importance of combining advanced AI systems with rigorous clinical oversight and independent safety evaluations like NOHARM."
Deployment and Security
Doximity's Clinical AI Suite, including Ask, has been deployed across more than 150 health systems, including eight of the nation's top 20 hospitals. The platform incorporates end-to-end encryption, role-based access controls, audit logging, and session isolation. These features are designed to help healthcare organizations deploy AI while maintaining enterprise-grade security and privacy standards.
How will this top ranking in the NOHARM benchmark influence Doximity's partnerships with the remaining top-tier hospitals not yet using the platform?
What are the plans for scaling the PeerCheck program as the platform expands to ensure the quality of physician oversight keeps pace with user growth?
Will Doximity seek to publish these results in peer-reviewed medical journals to further validate the clinical safety of its AI models?



























