Most U.S. doctors are quietly using this AI tool. Few patients know about it.

An AI-powered tool called OpenEvidence has become widely adopted among U.S. doctors, assisting in clinical decision-making, medical knowledge updates, and exam preparation. In April 2024 alone, the tool was used in approximately 27 million clinical encounters by around 65% of U.S. physicians—roughly 650,000 doctors—while another 1.2 million use it internationally. The platform provides tailored responses to medical queries, often replacing traditional peer consultations, and can generate patient discharge notes or study materials for licensing exams. Doctors across specialties and regions, from Hawaii to Maine, report relying on OpenEvidence for quick, evidence-based guidance. For example, a junior doctor in New Hampshire used it to confirm that a patient’s low potassium levels were a medication side effect, while a physician at the Pine Ridge Indian Health Service center in South Dakota consulted it to determine whether an X-ray or CT scan was needed to diagnose a suspected spine fracture. Both cases illustrate how the tool helps clinicians navigate complex medical scenarios efficiently. The rapid adoption has raised concerns among experts, including potential inaccuracies in AI responses and the risk of doctors becoming overly dependent on the tool. Dr. Anupam Jena, a Harvard professor and internal medicine physician, noted that 60% of OpenEvidence queries involve clinical decision-making, with doctors seeking treatment recommendations tailored to specific patient conditions. Jena is analyzing 90 million queries from 2024 to assess the tool’s impact on medical practice. Despite these concerns, many healthcare providers praise OpenEvidence for saving time and improving patient care. Dr. Jeremy Cauwels, chief medical officer at Sanford Health in Sioux Falls, South Dakota, described it as easy to adopt and highly functional, especially for rural clinicians managing large caseloads. The tool’s accessibility on mobile devices and its integration into daily workflows have contributed to its exponential growth in popularity. Critics argue that the lack of rigorous scientific studies on OpenEvidence’s patient outcomes and the possibility of AI-generated errors could pose risks. However, its ability to synthesize vast amounts of medical literature quickly has made it indispensable for doctors facing time constraints or complex cases. The debate over its benefits and drawbacks continues as AI tools reshape medical practice.