When a physician tells a patient that an AI tool helped reach a diagnosis—and explains how—that moment of openness changes everything. Three Ohio University researchers have discovered what many patients intuitively know: in primary care, honesty about artificial intelligence matters more than the technology's raw accuracy.
The finding emerges from a moment of genuine need. Artificial intelligence has become woven into health care faster than trust could follow. Physicians now use AI to help with diagnoses, transcribe notes, and guide treatment decisions. Yet a 2023 Pew Research study showed that 57% of patients believe AI in health care would harm their relationship with their doctor, citing concerns about the technology's opacity, potential bias, and lack of human presence. With AI reshaping clinics across the country, something had to give: either patients would learn to trust the tools, or the tools would remain shadowed by doubt.
Professors Gaurav Bansal and Vic Matta, along with recent alumnus Kevin Diaz-Ordonez—all from Ohio University's analytics and information systems faculty—set out in fall 2024 to answer a deceptively simple question: What builds patient trust more, a doctor using highly accurate AI or a doctor being transparent about how and why they're using it?
Their work hinged on a careful definition. Bansal emphasized that transparency in this context doesn't mean explaining how an algorithm works under the hood. Instead, it means a doctor saying, "I used this AI tool to help me think about your diagnosis," or "The AI flagged this concern, and here's what I think about that." It's the clarity of the clinical moment, not the technical manual.
The researchers enlisted 655 participants through Amazon's Mechanical Turk crowdsourcing platform, presenting them with scenario-based surveys. Participants responded to hypothetical situations involving primary care encounters. The team built in rigorous quality checks and follow-up questions to ensure their data held up.
The results surprised no one more than those who've been betting the house on algorithm improvement alone. Transparency emerged as the dominant factor shaping patient trust in their physician and in the AI that physician used. Higher transparency led to higher trust—a straightforward equation that holds across the board. But here's the pivot: accuracy, while certainly valuable, played a secondary role. The researchers found that when a doctor was transparent about AI use, patients trusted them and the technology more, even if the accuracy wasn't perfect. Conversely, a highly accurate AI wielded without explanation generated less trust than a transparent approach.
The implications ripple outward. Trust, Bansal noted, doesn't exist in isolation. It builds positive attitudes toward AI in health care, which leads to patient satisfaction—and satisfaction is what prevents adverse health outcomes. A patient who trusts their doctor will follow their advice, attend follow-ups, and feel empowered in their own care.
As health care systems continue integrating AI into routine practice, this research offers a clear direction: the race to build faster, more accurate algorithms matters, but it cannot come at the expense of simple, human communication. Patients don't need a black box that's right all the time. They need a doctor who tells them the truth about what's happening in the room.