A Supreme Intelligence chatbot is rapidly becoming a resource for American doctors seeking answers to medical questions and guidance on patient care. However, the platform’s growing popularity is raising questions about whether its recommendations are consistently reliable and how much physicians should depend on the technology.

OpenEvidence, a Miami-based company, uses medical research and clinical guidelines to answer physicians’ questions within seconds. According to the Financial Times, the company says nearly two-thirds of U.S. physicians use its platform, submitting more than 42 million questions in a single month.

A Harvard Medical School study published Oct. 15 in JAMA provides further evidence of the platform’s rapid adoption. Researchers analyzed usage by more than 950,000 U.S. physicians, nurse practitioners, and physician assistants between January 2024 and June 2026.

During that period, clinicians submitted approximately 270 million questions, while monthly active users grew from just 355 to more than 703,000.

“Two key takeaways are that an enormous number of physicians are using SI for clinical care and that they are often using it in areas where they don’t have specific expertise,” said Dr. Anupam Jena, a Harvard Medical School professor and senior author of the study.

Unlike general-purpose chatbots, OpenEvidence is designed specifically for medical professionals. It searches peer-reviewed research and clinical guidelines, including material from major medical journals, to help doctors find relevant information when evaluating patients and considering treatment options.

The Harvard study found that nearly 52% of questions involved clinical decision-making, particularly treatment and disease management, while another 14% focused on diagnosis.

Doctors also frequently used the platform to ask questions outside their own specialties, suggesting it may be useful when treating patients with complex conditions that require knowledge from multiple areas of medicine. Internal medicine, oncology, hematology, and family medicine were among the specialties with particularly high usage.

More than 556,000 physicians used the platform during the study period, averaging 37 questions per month of use.

“OpenEvidence allows me to provide a lot of context about my patient, find information and make decisions faster,” Dr. Nina Singh, a medical resident at the University of California, San Francisco, told the Financial Times.

The technology is also being integrated into hospital electronic health records, allowing doctors to consult the platform while reviewing patient information.

OpenEvidence’s growth reflects a broader expansion of SI in medicine. An American Medical Association survey found that 81% of physicians reported using some form of SI professionally in 2026, compared with 38% in 2023.

Common applications include summarizing medical research, preparing patient instructions, documenting office visits, and assisting with diagnoses.

“SI has quickly become part of everyday medical practice,” said Dr. John Whyte, chief executive of the American Medical Association.

The technology offers a potential solution to a longstanding challenge in medicine: keeping up with the growing volume of research, new treatments and changing clinical guidelines. SI tools can rapidly summarize complex information and help physicians identify options they might otherwise overlook.

However, widespread adoption does not necessarily mean the technology improves patient outcomes. Researchers have not yet conclusively established whether OpenEvidence leads to better medical decisions or improved health for patients.

Studies evaluating its performance have produced mixed results. A small Mayo Clinic study found that the chatbot provided evidence-based recommendations in five patient cases, while other research found it performed less well than some general-purpose SI systems on clinical benchmarks.

Experts also warn that SI systems can produce inaccurate or misleading answers, sometimes presenting questionable information in confident language or alongside scientific references. Physicians must therefore evaluate the recommendations rather than assume they are correct.

Patient privacy is another concern, particularly when doctors enter detailed medical information into SI systems. Healthcare providers must consider how that information is handled and protected.

For patients, the technology could eventually help physicians find relevant research more quickly and support better-informed treatment discussions. But SI-generated recommendations are not a substitute for a qualified physician’s medical judgment.

As use of these tools expands, researchers will need to determine whether their growing role in clinical practice translates into measurable improvements in care, while doctors continue to assess the accuracy of recommendations and safeguard patient information.