TL;DR
Medical organizations insist AI can only assist human doctors, but evidence shows fully autonomous AI already outperforms humans on some tasks, and resisting it only holds back progress.
The American Medical Association and other groups keep insisting that AI must only assist human doctors, never take charge. But the authors argue this position ignores the evidence and could deny millions of Americans better care.
Doctors misdiagnose too
Much of modern medicine is fairly algorithmic: ask questions, run tests, form a diagnosis, set a plan, adjust. Even following the guidelines, doctors make mistakes — up to 15 percent of US patients are misdiagnosed each year, leading to tens of thousands of preventable deaths.
Machines can already do well
In one study, experts rated Google's medical AI better than human doctors at drawing useful information out of people playing the patient. In a 2023 trial, patients with type 2 diabetes had their insulin adjusted by AI alone or by a doctor; the AI got patients to a stable dose faster, and patients stuck with it more reliably.
As for the human-AI teams the medical associations favour? A 2025 review of 52 studies found they did not beat AI alone or the best human clinicians.
'Robots lack empathy'? Doctors can be robotic too
A Gallup poll found 21 percent of people have turned to AI after being dismissed by a health-care provider, and in study after study patients rate AI as more emotionally supportive than human doctors.
The authors are not claiming AI is perfect. They just argue that instead of writing it off, the medical community should help figure out how to test, regulate and integrate it. Medicine once bled patients with dirty knives; it progressed then, and it can today.
Curated from high-quality sources, with concise summaries and key takeaways.
