A quiet revolution is reshaping American medicine, and most patients have no idea it is happening.
OpenEvidence AI, a clinical search platform built exclusively for physicians, has become one of the fastest-growing medical tools in the United States. Doctors use it daily to check research, weigh treatment options, prepare for board exams, and navigate complex patient questions during packed clinical schedules.
The numbers tell a striking story. OpenEvidence AI now reaches roughly 65% of U.S. doctors and touched nearly 27 million clinical encounters in April alone, according to company figures. Close to 650,000 American physicians actively use the platform. Another 1.2 million users log in from outside the country.
This surge fits a broader transformation sweeping health care. The American Medical Association reported in March 2026 that 81% of physicians now use machine-learning tools professionally. That figure has more than doubled since 2023. Doctors most commonly apply these tools to summarize research, generate discharge instructions, document visits, and produce chart summaries.
A clinical habit spreads fast
OpenEvidence AI functions like a research assistant embedded inside a doctor’s workflow. A physician types a clinical question in plain language. The platform scans medical literature, treatment guidelines, and licensed journal content, then delivers a concise, cited answer within seconds.
That speed drives adoption.
Doctors today face an avalanche of new studies. They also manage overflowing patient schedules, tight appointment windows, and heavy documentation requirements. A tool that compresses a 20-minute literature search into seconds feels less like a gadget and more like a lifeline.
Dr. Anupam Jena, an internal medicine physician at Massachusetts General Hospital and a Harvard health policy professor, put it plainly.
“Everyone is using it,” Jena said.
He noted that most queries center on real bedside decisions. Physicians want to know the right treatment for a specific patient dealing with a specific condition, medication history, or risk profile.
Investors take notice
OpenEvidence AI has also drawn serious financial firepower. Reuters reported in January 2026 that the company raised $250 million at a $12 billion valuation. That valuation doubled in just three months. The company disclosed that its platform had already supported about 18 million clinical consultations from verified U.S. physicians by December 2025.
The platform has moved to deepen its content by striking deals with major medical publishers. The JAMA Network announced a multiyear content agreement with OpenEvidence in June 2025, covering JAMA, JAMA Network Open, and several specialty journals. The American Diabetes Association followed in December 2025, partnering with OpenEvidence to make its Standards of Care in Diabetes more accessible and searchable for clinicians.
Safety questions grow louder
Yet the same speed that attracts doctors also alarms health experts.
OpenEvidence AI maintains that its tool supports physician judgment rather than replacing it. That distinction carries real weight. Clinical decisions carry direct patient risk. A wrong answer can delay a diagnosis, prompt an unnecessary test, or lead to a harmful medication choice.
Some physicians told NBC News they cross-check sources when an answer surprises them. Others rely on the tool for routine queries but stay cautious when cases turn unusual or rare.
The AMA survey revealed deep concern. Eighty-eight percent of physicians reported worrying about skill erosion linked to health-focused machine learning tools. The same share called for strong safety and effectiveness validation. Eighty-six percent identified data privacy as a critical condition for broader adoption.
Hospitals have their own worries. Clinicians using personal devices to access unvetted tools, without system oversight, represent what insiders call “shadow AI.” It refers to unauthorized clinical AI use invisible to hospital administrators and compliance teams.
Dr. Girish Nadkarni, who leads AI work at Mount Sinai Health System, acknowledged that hospitals cannot ignore the trend. Mount Sinai responded in March 2026 by integrating OpenEvidence AI directly into its electronic health record system for doctors, nurses, and pharmacists.
Competition arrives in the exam room
OpenEvidence AI no longer operates in an empty field.
OpenAI launched ChatGPT for Clinicians in April 2026, opening it to verified U.S. physicians, nurse practitioners, physician assistants, and pharmacists. The company said the tool helps clinicians with research, documentation, and clinical decision support, with cited responses drawn from trusted medical sources.
That launch signals where the industry is heading. The next race centers on verified clinician access, licensed medical content, privacy controls, citation accuracy, and deep hospital integration.
For patients, one question stands above the rest. Should doctors disclose when a clinical AI tool shapes their care?
Many physicians view OpenEvidence AI as another reference resource, no different from a textbook or medical database. Others argue that patients have a right to know when digital systems influence clinical reasoning.
The answer likely hinges on how doctors use it. A quick background check may not require disclosure. A tool that drives a diagnosis or shapes a treatment plan raises a more serious question.
OpenEvidence AI has already changed how hundreds of thousands of doctors search for answers at the bedside. Now, hospitals, regulators, and patients must decide how transparent that change needs to be.
Do you think doctors should tell patients when they use OpenEvidence AI or a similar clinical tool during their care? Please drop your thoughts in the comments below.