Advances in medicine have long relied on a powerful safeguard: peer review. Before a therapy is accepted or trial results are trusted, experts interrogate the evidence. That human process, imperfect yet indispensable, has upheld scientific rigor and protected patients for decades.
Today, as generative AI enters daily clinical practice, that safeguard is often missing. AI can synthesize vast amounts of information, but it cannot replace clinical judgment, context, or lived experience. An important study that stress-tested current large language models revealed serious deficiencies across key clinical benchmarks and concluded they are not ready for use.
This is why former U.S. Surgeon General Dr. Regina Benjamin and I have agreed to serve as co-editors in chief and launch PeerCheck, a physician-led initiative to bring true medical peer review into the AI era.
The PeerCheck Model
PeerCheck will embed physician review into DoxGPT, Doximity’s evidence-based AI tool. Our immediate goal is to bring more trust and transparency to AI outputs by allowing physicians to evaluate and improve AI-generated answers for accuracy, evidence strength, and potential bias. Long term, our aim is for PeerCheck to serve as a model for how medical AI should be built and governed.
The Work Ahead: A Call to My Colleagues
The effort ahead is ambitious. Building an Editorial Board, defining specialty oversight, and developing robust bias-mitigation processes are essential if medical AI is to earn clinicians’ trust and evolve responsibly.
Regina and I will convene our inaugural AI Editorial Board meeting in San Francisco in March. We’re seeking physicians to help review content and guide this framework. If you’re interested in participating, please let us know below.
Together, we can build AI systems worthy of our profession and our patients’ trust.
Eric J. Topol, MD