Medical experts and government advisers are debating whether artificial intelligence will eventually replace human physicians in diagnostics and treatment planning.
The shift toward AI-driven healthcare matters because it could fundamentally change how patients receive care, potentially increasing diagnostic speed while introducing new risks regarding accuracy and regulation.
Recent studies suggest that large language models can perform as well as or better than physicians in certain diagnostic cases [3, 4]. Some reports indicate that AI has specifically outperformed emergency room doctors in diagnostic scenarios, suggesting the technology could take over specific physician tasks [3].
However, the transition is not without significant errors. Dr. Nitish Naik said that AI invented a disease that never existed [5]. This highlights the risk of "hallucinations," where AI generates plausible but entirely false medical information.
Despite these errors, some officials are preparing for a future where AI operates independently. Amy Gleason, an AI adviser for the U.S. Department of Health and Human Services, said that discussions are underway to create a regulatory pathway for independent AI physicians [6]. Gleason said this potential transition is like the decades-long process that moved self-driving cars from test tracks to city streets [6].
Not all healthcare leaders agree with this trajectory. Some health system leaders have described the concept of AI physicians as "absurd" [6]. These critics argue that the complexity of human health, and the necessity of clinical judgment, make the total replacement of doctors unfeasible.
The debate currently centers on whether AI will serve as a tool to assist doctors—augmenting their ability to read medical imaging and plan treatments—or if it will eventually function as a standalone provider [3, 4, 5].
“"AI performed as well or better than physicians in the new study."”
The tension between AI's diagnostic speed and its tendency to hallucinate creates a critical gap in patient safety. While regulatory frameworks for 'AI physicians' suggest a move toward automation, the current lack of reliability in complex cases means AI is more likely to remain a supportive tool for human clinicians rather than a replacement in the near term.
