AI and the Future of Medicine

Lakshmi Krishnan, a physician and researcher, recently discussed the impact of artificial intelligence (AI) on medical practice in an editorial published in Stat. A recent study revealed that ChatGpt-4 achieved a diagnostic reasoning score of 90%, surpassing the 76% of physicians, even when they used the chatbot for assistance. This result raises critical questions about the nature of medical thinking and the potential for AI to replace physicians. Krishnan emphasizes that the adoption of new technologies has historically raised similar concerns. Technologies such as the stethoscope and X-rays were initially seen as threats to the doctor-patient relationship, while the electronic health record, introduced to improve efficiency, led to new complications. However, every innovation has found its place as a useful tool within the medical system rather than as a substitute for professionals. Diagnosis, an ancient practice, became crucial with the development of pathological anatomy. Physicians are often viewed as detectives gathering clues to formulate a diagnosis, but this view is limited. Diagnosis involves not only identifying patterns but also interpreting evidence and narratives, considering the patient's lived experience and context. AI can recognize patterns but cannot assess the significance of lived experience. Additionally, diagnostic inequalities reveal that errors stem not only from a lack of knowledge but also from systemic and access issues. AI algorithms can even amplify existing inequalities in the healthcare system. AI does not replace medical reasoning but integrates it by offering new insights. ChatGpt has succeeded not by mimicking physicians but by using pattern recognition to apply medical knowledge. AI's ability to explain its reasoning can stimulate and improve medical thinking. However, many physicians use AI in a limited way, not fully harnessing its analytical potential. It is essential for medical training to evolve, focusing less on the mere acquisition of biomedical knowledge and more on developing the ability to interpret contexts and narratives. Physicians must become attentive readers of scientific literature and human stories, capable of articulating their reasoning and interpreting biological, social, and cultural contexts. AI should thus be seen as an ally for physicians, helping them reflect on their approach and the inequalities in the healthcare system. The goal is to effectively integrate AI into the medical profession.