Should AI Be Allowed to Make Medical Decisions?

• FDA-approved AI diagnostic tools already operate autonomously for specific conditions. IDx-DR reads retinal scans without any physician review, and the FDA's 2018 authorization was grounded in clinical evidence that autonomous AI was safer than the available alternative, which was no screening at all. Extending this model to other well-validated conditions follows the same regulatory logic that has already been applied.
• Studies in Nature Medicine and The Lancet have shown AI systems outperforming average board-certified specialists in targeted tasks: detecting breast cancer on mammograms, identifying diabetic eye disease, and classifying skin lesions. The relevant comparison is not AI versus the best physician in the world; it is AI versus the physician who is actually available, which in most of the world means no physician at all.
• Requiring physician sign-off on every AI recommendation reinstates the scarcity bottleneck that AI was built to eliminate. If an AI correctly diagnoses tuberculosis from a chest X-ray in seconds, mandating that a physician review every result before treatment delays care by hours in contexts where physician time is the binding constraint, not diagnostic accuracy.

Debater: To be announced

• AI medical systems fail in ways that individual physician errors do not. A physician who misdiagnoses a patient makes an individual error; an AI that misdiagnoses due to training data bias produces the same error for every patient with similar demographic characteristics. When that bias favors lighter skin tones, as documented in dermatology AI systems, the error is not random; it is structural and discriminatory, and it scales.
• No existing legal framework assigns liability when an autonomous AI kills a patient. The device manufacturer disclaims clinical responsibility, the hospital claims the device is approved, and the patient has no physician to hold accountable. Medicine requires accountability, and accountability requires a human decision-maker present at the point of consequence. Autonomy without liability is a gap the law has not closed.
• Physicians integrate information that does not appear in structured data: the patient's emotional state, whether they will adhere to a treatment, the domestic circumstances that make a specific intervention impractical. Diagnosis is a synthesis of everything the patient communicates and everything they cannot articulate. Removing the physician removes that synthesis, and the cases where it matters most are precisely the cases AI is least equipped to handle.

Debater: To be announced