AI in Clinical Practice vs. Academic Research
Artificial intelligence is increasingly capable of performing specialist-level clinical tasks, including diagnosis, treatment planning, prescription management, and multi-disciplinary patient follow-up. However, AI remains largely a tool in academic and medical research contexts, where hypothesis generation, experimental design, and novel knowledge production are still human-led. Medical schools and research institutions continue to train Medical Doctors who perform medical human body related research, roles that AI cannot fully replace in the near future Thus, while AI can assume many specialist clinical functions, it cannot yet substitute for the combined expertise of academic research scientists in core biomedical fields such as Biochemistry, Biophysics, Physiology. However, in academic research, AI is already successful and in the 20 years far future it will take over at least 70% of this field.
In the very near term, artificial intelligence is poised to assume the roles traditionally held by clinical specialists across a wide range of disciplines—including radiology, internal medicine, cardiology, pulmonology, neurology, oncology, pediatrics, gastroenterology, nephrology, dermatology, psychiatry, infectious diseases, hematology, ophthalmology, rheumatology, immunology, geriatrics, obstetrics and gynecology, anesthesiology, emergency medicine, critical care, urology, otolaryngology, physical medicine and rehabilitation, pain management, and others—handling diagnosis, treatment planning, prescription management, and multi-disciplinary patient follow-up. While foundational medical education and biomedical research remain human-led, AI will effectively take over practical clinical decision-making tasks.
Actually, it is very easy to know what AI will dominate. AI will handle—and already is handling—everything (except computer engineering work) done at a desk or any physician work that would be done sitting at a table. In other words, AI is now performing tasks that previously required highly trained humans.
For example, AI is taking over tasks from mid-level educated healthcare workers, but it is also replacing the desk-based work of highly trained, educated specialist clinicians. Essentially, any task that can be done at a desk is at risk of being fully automated by AI.
AI will take over the work of a top-tier surgeon—whose skills appear in one-in-a-million cases—relatively soon. While it can also take over the tasks of an ordinary nurse, doing so is much more expensive for robotic AI, so there will still be a human need for moderately trained healthcare workers like nurses (Human nurses will be needed at least until 2035).
In the future, engineering will remain a profession that relies on human intelligence, as it requires designing, maintaining, and improving AI-driven systems. In contrast, desk-based professions that demand high-level reasoning—such as law, clinical decision-making, and certain administrative roles—will increasingly be performed by AI. This shift may lead to a decline, or even extinction, of human-led expertise in these domains, as AI assumes tasks once requiring superior human cognition.
For example, AI platforms such as Replit AI and Bolt AI—particularly Replit—are already performing master or above level engineering tasks and have replaced many engineers. As a result, engineers at certain levels are losing jobs, even as new positions are created by AI itself.
