Recently, I learned that access to a clinical artificial intelligence platform had been restricted within my training environment. My first reaction was not anger. It was curiosity.
The decision may have been entirely reasonable. Health care institutions have legitimate concerns about privacy, security, liability, accuracy, and governance. Those concerns deserve thoughtful consideration. But the experience left me wondering about something much bigger than any individual hospital or software platform. If physicians are expected to practice evidence-based medicine, what responsibility do health care systems have to ensure physicians can efficiently access evidence?
As physicians, we often talk about evidence-based medicine as if evidence itself is the primary challenge. It is not. The challenge increasingly is access.
Every day, I see patients with complex questions that deserve thoughtful, evidence-informed answers. Like most physicians, I want to know the latest data. I want to know what guidelines say. I want to understand where controversy exists and where consensus has emerged.
Real physician voices, twice a week
Free, and one click to unsubscribe.
But modern medicine has a problem. The volume of medical knowledge has grown beyond what any individual physician can fully absorb. Thousands of studies are published every week. Guidelines are updated. New technologies emerge. Clinical recommendations evolve. At the same time, physicians are managing packed clinic schedules, documentation requirements, prior authorizations, quality metrics, inbox messages, and electronic health records.
The result is a paradox. We have more medical knowledge than at any point in history, yet finding the right information at the right moment often feels harder than ever.
This issue is becoming especially relevant as major professional societies begin exploring partnerships with AI-enabled platforms to improve access to evidence-based guidance. In women’s health, where clinicians routinely manage rapidly evolving questions across obstetrics, gynecology, reproductive endocrinology, menopause, contraception, surgery, and preventive care, the ability to access reliable information efficiently is not a luxury. It is central to clinical care.
Historically, medicine has solved information problems through innovation. Medical libraries gave way to digital databases. Paper journals became searchable online archives. Search engines transformed literature retrieval. Clinical reference tools brought evidence closer to the bedside. Each innovation was met with skepticism. Many eventually became indispensable.
Imagine telling a modern physician that the most reliable way to answer a clinical question is to leave clinic, walk to a medical library, locate a textbook, search through an index, and manually review references. The idea seems absurd. Yet every generation has struggled with technologies that challenge established ways of learning and working.
We do not say someone has never truly experienced Europe because they traveled there by airplane instead of by boat. We do not insist that navigation was better before GPS because previous generations relied on paper maps. The value lies in reaching the destination, not preserving the inefficiencies of the journey.
Artificial intelligence may represent the next chapter in this evolution. Much of the public discussion surrounding AI focuses on whether these systems will replace physicians. That conversation misses how many clinicians are actually using them.
Most physicians are not asking AI to make diagnoses independently. They are not asking it to replace clinical judgment. They are asking it to help them find information. They are asking it to help identify relevant studies, locate guidelines, synthesize evidence, and navigate an increasingly complex medical literature. In other words, they are asking it to do what every information technology has attempted to do for generations: Reduce friction between knowledge and application.
Critics correctly point out that AI can make mistakes. So can textbooks. So can review articles. So can experts. Medicine has never functioned by blindly accepting information from any source. Physicians routinely verify recommendations, evaluate evidence, compare sources, and exercise judgment. The solution to imperfect tools has never been to abandon them altogether. The solution has been to understand their strengths, recognize their limitations, and use them responsibly.
What concerns me most is not the possibility that physicians will use AI. That has already happened. What concerns me is the possibility that health care institutions will focus so heavily on the risks of information technology that they overlook the risks of information inaccessibility.
When physicians cannot efficiently access trusted information, patients do not benefit. When evidence becomes difficult to find, evidence-based medicine becomes more difficult to practice. When clinicians spend more time searching for information than applying it, everyone loses.
The future of medicine should not be a choice between uncritical adoption and blanket prohibition. It should be a conversation about responsible implementation. How do we evaluate new tools? How do we establish safeguards? How do we maintain transparency? How do we educate physicians about appropriate use? And perhaps most importantly, how do we ensure that the best available evidence reaches physicians when patients need it most?
For decades, medicine’s greatest challenge was generating knowledge. Today, our greatest challenge may be ensuring that knowledge is accessible. Patients deserve care informed by the best available evidence. Physicians deserve tools that help them access that evidence efficiently, responsibly, and transparently. Evidence-based medicine requires evidence. Increasingly, it also requires access.
Nicholas M. Stansbury is an obstetrician-gynecologist and reproductive endocrinology and infertility fellow.
