By Ali Araghi, MD

Seventy years ago, T.R. Harrison, who was the first editor of Harrison’s Principles of Internal Medicine, established an elegant standard of being a physician: “No greater opportunity, responsibility, or obligation can fall to the lot of a human being than to become a physician. In the care of the suffering he needs technical skill, scientific knowledge, and human understanding. He who uses these with courage, with humility, and with wisdom will provide a unique service for his fellow man, and will build an enduring edifice of character within himself. The physician should ask of his destiny no more than this; he should be content with no less.”

After 70 years, I rarely see a medical student or even an internal medicine resident in the US who has heard of this book, let alone know Harrison and his famous quote. After 70 years, the art and science of making rapport with patients, taking a thorough history with empathic listening, feeling their pulse, listening to their hearts and lungs and constructing a thoughtful differential diagnosis, diagnostic, and care plan are all forgotten. We all start rounds at computer stations, checking numbers and imaging reports, focused mainly on timely documentation and coding rather than seeing patients first and doing the paperwork later. Bedside rounds and discussions have been reduced to brief, closed-door reviews of labs and imaging, with a light, fast encounter with patients. This cultural shift is one of the least considered drivers of health care costs and distrust between patients and physicians (not providers!). Indeed, the commoditization of medicine through the corporatization of health care sadly but truly degenerated physicians into providers. Right at this juncture, AI enters the field of medicine. There is no doubt that wise and proper use of AI will have a huge impact on improving quality of care, reducing waste, and enhancing medical education beyond imagination, but the secret lies in two words: “wise and Proper”. As Yuval Harari said in his 2026 Oxford Tanner Lecture titled “AI bureaucrats, AI religions and AI boyfriends”, it will be a huge mistake if we think that AI is a tool. It is an agent. It is worth summarizing his main points here because they’re highly relevant to practicing any profession, including medicine, starting now.

Harari’s distinction is fundamental: a tool executes a purpose chosen by its user, whereas an agent can interpret circumstances, make intermediate decisions, learn, and pursue outcomes that its designers did not explicitly specify. AI need not be conscious to possess this practical agency. In medicine, therefore, it should not be understood merely as a faster calculator or reference book. An AI system that prioritizes patients, frames diagnostic possibilities, drafts clinical judgments, or shapes communication is participating in decisions. The central question is no longer only whether the technology is accurate, but also who defines its goals, supervises its choices, and remains accountable when its recommendations influence care.

Harari further argues that language is civilization’s operating system because laws, institutions, money, trust, and shared meaning are all constructed through words. Medicine also depends on this cognitive domain: symptoms become histories, observations become diagnoses, and clinical language shapes both institutional decisions and a patient’s understanding of illness. Because AI can generate and manipulate words at extraordinary speed and scale, it may increasingly mediate what clinicians notice, how they reason, and what patients are told. If physicians surrender that domain uncritically, they risk retaining formal authority while losing intellectual and moral agency and relevancy. The proper response is neither rejection nor passive adoption, but disciplined oversight that uses AI to strengthen scientific judgment and restore human attention, rather than replace either. With the current dominance of this data-centered approach to address a human disease just as abnormal data rather a suffering human as whole, it is no surprise to hear statements and interviews from people like Elon Musk, who has argued that AI already outperforms most physicians and will improve further (Becker’s Hospital Review, August 8, 2025); predicted that Tesla’s Optimus robots could surpass the best human surgeons at scale within three years, making traditional medical training seem “pointless” (The Independent, January 8, 2026); and suggested that the doctor of the future will be an AI system, perhaps one not yet widely known (PRIMETIMER, July 26, 2026). What this ideology terribly misses is the power of human connection in healing another human being. Healing is a far deeper concept than merely regressing abnormal data toward expected ranges. Even if an AI agent can initiate an emotional and supportive conversation with a patient, it cannot replace the genuine, warm touch of an examiner’s hand or the kind look of an empathic physician to her patient. The power of healing emanates from the human soul and penetrates another human soul. To safeguard our profession against this evil ideology, we need to resurrect the art of medicine in our daily practice routines and teach our fellow students, nurses, and physicians the wisdom of being a proud healer, not a provider of diagnostic and therapeutic technologies like a senseless robot.