Who’s the Smartest Person in the Exam Room?

Traditionally, this question had a clear answer: the physician. With years of education, clinical experience, and mastery of medical science, the doctor has long been the intellectual authority in healthcare. But the dynamics of the exam room are shifting. Increasingly, the smartest person might not always be the physician. It could be the parent of a child with a rare disease, whose lived experience has made them a de facto expert. It might even be the patient, empowered by hours of research and insights gleaned from the internet.

Today, however, the answer might surprise you. The smartest presence in the room could very well be the computer. With the rise of large language models (LLMs), the exam room now includes an unprecedented repository of knowledge and computational power. These systems process vast amounts of medical information faster than any human, creating new opportunities—but also introducing new tensions—about authority, collaboration, and trust in healthcare.

The Emerging Triad: Patient, Clinician, and AI

The integration of LLMs into healthcare has transformed the exam room into a three-way conversation. Each participant brings unique strengths and perspectives. The patient, armed with AI tools, often arrives better informed and prepared than ever before, ready to engage in their care. The clinician, meanwhile, uses AI to enhance their practice, drawing on tools that analyze patient histories, identify rare conditions, and recommend treatments with remarkable speed. And then there is the AI itself, a neutral and tireless presence offering access to an expansive universe of medical knowledge.

This triad—the patient, the clinician, and the AI—has the potential to create a richer and more informed dialogue. It promises to refine how diagnoses are made, treatment plans are crafted, and information is shared. However, it also introduces a new dynamic, one that demands careful navigation. When the AI holds the most raw knowledge in the room, who becomes the ultimate authority? And how do we resolve the inevitable conflicts that arise between machine-driven insights and human expertise?

Tensions in the Exam Room

This new dynamic in the exam room is both exciting and challenging. On one hand, LLMs empower patients to better articulate their symptoms and advocate for their care. On the other hand, this empowerment can sometimes verge into overconfidence. A patient, buoyed by an AI-generated suggestion of a rare diagnosis, may push for unnecessary tests or treatments. For the clinician, navigating this enthusiasm requires a delicate balance of empathy and expertise. And in an interesting twist, it might be AI—or even the patient—who hears that fabled zebra’s hoofbeats.

Clinicians, too, must grapple with the dual role of AI as both ally and challenger. While LLMs enhance diagnostic capabilities and save valuable time, they can also threaten the perception of the clinician’s authority. Patients may question, “Why should I trust you over what the AI suggests?” This tension forces clinicians to reaffirm their value not just as experts but as interpreters of both data and humanity. Unlike AI, clinicians bring intuition, experience, and empathy to the equation—qualities that are critical in contextualizing and personalizing care.

Efficiency, another hallmark of AI-enhanced communication, is a double-edged sword. Streamlining the exchange of information between patient and clinician can lead to faster diagnoses and treatment decisions. However, speed isn’t always an asset in healthcare. Patients often need time to process complex information and emotionally engage with their care. Clinicians must navigate the subtleties of a patient’s narrative, which cannot always be captured in the algorithmic precision of an AI model. Nevertheless, the tailored articulations of LLMs offer a tremendous opportunity to elevate communication to a new and more engaging level.

Still, one of the most pronounced tensions arises when patient-generated AI insights conflict with clinician-generated AI recommendations. In these moments, the exam room can feel like a battleground of algorithms. The question becomes not just “Who is right?” but “How do we resolve this conflict in a way that maintains trust and collaboration?” My sense is that this might get worse before it gets better, particularly when these conflicts arise in real time.

Beyond the Triad: The Broader Ecosystem

While the triad of patient, clinician, and AI forms the core of the modern healthcare interaction, it is not the entire picture. The exam room exists within a far more complex ecosystem. Payors, regulatory bodies, healthcare systems, and social perception all play varied roles in shaping how care is delivered. The integration of LLMs into this broader framework raises additional challenges.

For example, insurers may begin to rely on AI-driven insights to approve or deny treatments, creating potential conflicts between patient needs, clinician judgment, and algorithmic decision-making. Healthcare systems must grapple with ensuring equitable access to LLMs, preventing disparities between patients who can effectively use these tools and those who cannot. And when errors inevitably occur—an incorrect AI recommendation or a misinterpreted output—questions of accountability will come to the fore. Who is responsible: the clinician, the AI developers, or the systems that integrated these tools?

Amidst these complexities, the triad remains the focal point. It is in the exam room, where patient stories meet clinician expertise and AI’s computational power, that the future of care is being shaped.

The Path (with Bumps) Forward

So, who’s the smartest person in the exam room? The answer is no longer straightforward. Intelligence in healthcare is no longer about who holds the most knowledge; it is about how that knowledge is shared, interpreted, and applied. In this new era, the smartest “presence” is not any single participant—it is the evolving conversation itself.

The triad of engagement—patient, clinician, and AI—has the potential to redefine healthcare, making it more informed, precise, and empathetic. But realizing this potential requires intentional effort. Clinicians must embrace their roles as mediators, patients must critically engage with AI insights, and AI must remain a tool in service of the human relationship.

Looking beyond this triad to the broader orchestration of healthcare, one truth remains clear: the heart of medicine will always be the connection between people. No amount of computational power can replace the trust, empathy, and understanding that make healthcare not just a science, but an art.