Dr. Emanuel proposed that sufficiently capable autonomous AI might eventually be considered not merely a tool, but a practitioner, and therefore might require certification and licensure. He discussed potential approaches for establishing standards to evaluate and certify such systems.
He also challenged one of medicine’s traditional distinctions between physicians and machines: the humanity, empathy, and “art of medicine” associated with being a physician. Although these qualities are frequently cited as uniquely human strengths, he questioned whether the profession consistently demonstrates them in practice.
Throughout the meeting, this discussion repeatedly returned to a fundamental question: What Does it Mean to Be a Doctor?
Autonomous Tasks Versus Medical Practice
Dr. John Whyte, CEO and Executive Vice President of the AMA, offered an important counterpoint. He emphasized the distinction between an AI system autonomously performing specific medical tasks and the broader practice of medicine.
Medical practice exists within a professional and social framework in which someone is ultimately accountable for patient care. As AI assumes greater autonomy, however, determining where that responsibility resides becomes increasingly complex. If an autonomous system makes a clinical recommendation that results in harm, is responsibility borne by the physician, healthcare organization, AI developer, vendor, or another entity?
Dr. Humayun Chaudhry, President and CEO of the Federation of State Medical Boards, emphasized medical licensure and professional standards. The discussion underscored that the future of AI in medicine will be shaped not only by technological capability, but also by how we define competency, accountability, professional responsibility, and patient protection.
Several participants also suggested that advances in AI could challenge physicians to become better practitioners. As AI becomes increasingly capable of performing diagnostic, informational, documentation, and other cognitive tasks, physicians may need to place even greater emphasis on judgment, communication, empathy, trust, and the physician-patient relationship.
The Physician’s Role in the Next 5 to 15 Years
One panel considered how the physician’s role may evolve over the next five to fifteen years.
Although AI technology is advancing rapidly, healthcare systems may change more slowly. Medicine remains dependent upon hospitals and medical offices, legacy information systems, reimbursement structures, regulatory requirements, and established clinical workflows. These existing systems and processes may significantly slow AI implementation.
Single-payer healthcare was also discussed, particularly in relation to whether greater centralization could facilitate the scaling and implementation of AI technologies.
Are Existing Regulatory Structures Ready?
Participants also questioned whether current regulatory structures are prepared for increasingly autonomous AI.
One panelist characterized state medical boards as anachronistic, arguing that their basic structures were developed for a very different era of medicine. Questions were also raised about whether the U.S. Food and Drug Administration is adequately structured to regulate rapidly evolving AI systems.
These discussions raised an important issue for organized medicine: Can increasingly autonomous AI be incorporated into existing regulatory frameworks, or will new approaches to certification, licensure, oversight, and accountability be required?
The Patient Perspective
One panel provided an especially important perspective. A clinician described having a “love-hate relationship” with AI, reflecting both enthusiasm for its potential and concern about its risks.
A patient on the panel noted that the healthcare system frequently fails to meet patients’ needs. Consequently, some patients are turning directly to AI for information and guidance, not simply because it is convenient, but sometimes because they are frustrated with the healthcare system or desperate for answers.
This observation deserves our attention. As patients increasingly turn to AI, organized medicine should ask not only whether they should rely on these systems, but also why they feel the need to do so.
What Does it Mean to Be a Doctor?
The meeting concluded with a reception, where leaders, panelists, and participants continued many of these conversations informally.
The discussion in Washington demonstrated that the debate surrounding AI in medicine is moving beyond whether physicians should use AI. The questions now include what decisions AI should be permitted to make autonomously, how competency should be assessed, who bears responsibility when AI makes an error, and whether sufficiently advanced AI systems should ever be certified or licensed.
As AI becomes capable of performing more functions traditionally associated with physicians, medicine will need to define more clearly the distinction between performing medical tasks and practicing medicine.
For organized medicine, the challenge is not simply to respond to advances in AI, but to help establish the standards, responsibilities, and professional values that should govern its use.
The question that repeatedly surfaced in Washington may therefore be one of the defining questions facing our profession: What does it mean to be a doctor in the age of artificial intelligence?
