A growing number of doctors across the United States are turning to artificial intelligence tools to help guide clinical decisions, review medical research and manage patient care — often without patients realizing it.
One AI platform in particular, called OpenEvidence, has rapidly become one of the most widely used medical AI tools in the country, according to a report by NBC News. The company claims nearly 65% of US physicians used the platform across almost 27 million clinical encounters in April alone.
The rise of the technology is reshaping how doctors access medical knowledge, diagnose conditions and make treatment decisions, while also sparking debate over patient safety, privacy and doctors becoming overly reliant on AI systems.
What is OpenEvidence?
OpenEvidence is an AI-powered medical search and decision-support platform designed specifically for healthcare professionals.
Unlike general-purpose chatbots, the platform searches peer-reviewed medical journals, treatment guidelines and clinical databases to answer doctors’ questions in real time.
NBC News reported that physicians use it to:
-Review treatment options
-Double-check diagnoses
-Identify medication side effects
-Summarise medical research
-Write discharge notes
-Prepare for medical licensing exams
Doctors access the service through a website or mobile app after verifying their professional credentials using a government-issued healthcare identification number.
“Our commitment is that core OpenEvidence will always be free for users,” OpenEvidence CEO Daniel Nadler told NBC News.
Rapid adoption among doctors
Doctors interviewed by NBC News described OpenEvidence as becoming almost ubiquitous in hospitals and clinics.
“Everyone is using it,” said Dr. Anupam Jena, a physician at Massachusetts General Hospital and professor at Harvard.
“Its growth really has been exponential,” he added.
According to the news outlet, OpenEvidence is now used by roughly 650,000 physicians in the United States and another 1.2 million internationally.
Many doctors said the tool helps them quickly answer specialised medical questions that would otherwise require lengthy searches through journals or traditional reference systems.
Dr. Paul Sax of Brigham and Women’s Hospital told the news outlet that OpenEvidence’s search abilities “border on miraculous.”
“The process of searching for answers is frictionless,” Sax said.
How doctors are using the tool
NBC News spoke to doctors across multiple specialties who described using OpenEvidence during real patient encounters.
One physician reportedly used it to determine whether a patient’s dangerously low potassium levels were a normal side effect of medication.
Another doctor used the platform to confirm that a CT scan — not an X-ray — was needed to properly diagnose a spinal fracture.
The system generates answers by summarising published medical research and linking directly to peer-reviewed studies and clinical guidelines.
Many clinicians said the tool saves significant time compared with traditional medical reference platforms such as UpToDate.
Concerns about AI errors and “hallucinations”
Despite its popularity, experts warned NBC News that OpenEvidence is not infallible.
Like other AI systems, it can occasionally produce inaccurate information, overstate conclusions from limited studies or generate incomplete answers.
Some doctors interviewed by said they routinely verify the studies cited by the AI before acting on its recommendations.
“I usually click through to the referenced papers,” emergency physician Dr. Kassel Galaty was quoted as saying.
Others expressed concern that younger doctors and medical students could become too dependent on AI tools and lose critical diagnostic reasoning skills over time.
Privacy and data concerns
OpenEvidence says it complies with HIPAA, the federal US health privacy law, and has safeguards for handling protected patient information.
However, some hospital systems remain cautious.
NBC News reported that MaineHealth currently advises doctors not to enter protected health information into the platform.
Several physicians said they use patient details such as age, sex and medical history in searches, while avoiding names or direct identifiers.
