The stack of charts waiting after a long call shift is not a minor inconvenience. For many physicians, it is the thing that finally breaks them. Elizabeth Garchar, an OB-GYN specializing in maternal fetal medicine and perinatology, knows that terrain well, and when she talks about artificial intelligence in medicine, she does not start with diagnostics or clinical decision support. She starts with the paperwork.
Garchar is most energized by the prospect of AI that can streamline charting, audit billing for errors, and absorb the administrative friction that consumes hours physicians could spend on patients. "I am very excited about the possibility of AI helping to streamline charting, to make billing less of a nightmare by double checking my work and really making me focus on the medicine by doing all the paperwork it possibly can," she said. The throughline in her thinking is physician burnout, a crisis she sees AI as genuinely equipped to address, precisely because the worst of that burden is procedural rather than clinical.
But her enthusiasm has a hard edge. The same efficiency that could liberate physicians from rote tasks could, if applied carelessly, erode the habits of mind that make medicine work. Her concern is not that AI will malfunction. It is that physicians will stop questioning it.
"It is really important to keep our creativity as we practice medicine and not just do things automatically because it's what we did last time."
That warning carries particular weight in a specialty like maternal fetal medicine, where pregnancies that look routine can turn quickly, and where protocol-driven thinking has real limits. Garchar draws a distinction between using AI to support clinical reasoning and using it to replace that reasoning altogether. Following an AI-suggested pathway without interrogating it, she suggests, is not efficiency. It is a different kind of negligence, one that arrives wearing the clothes of best practice.
What the broader conversation about AI in medicine often misses, she argues, is the question of what gets lost when physicians stop doing the hard cognitive work themselves. Checklists and in-place protocols have their place. Research and clinical judgment have theirs. The risk is not that the machine enters the room. It is that the physician, relieved of paperwork and nudged by algorithms, gradually stops asking whether the protocol actually fits the patient in front of them.
