A Nameless Paper Says AI Beats Doctors. That's Not How Evidence Works.

A recent paper — unattributed, unjournal'd, undated — argues that AI is often better at doctoring than doctors. The claim has prompted visible anxiety among physicians about their professional role. The article frames this as a labor-market reckoning for medicine. None of that framing can be evaluated, because the paper producing it has no visible authors, no institution, no publication date, and no accessible methodology.

Strip the absent paper away and what remains is a bare assertion: "AI is often better at doctoring than doctors." That assertion functions as a marketing claim — structurally designed to land a conclusion rather than demonstrate one. There is no population to interrogate, no diagnostic task defined, no comparison protocol to scrutinize. Encouraging it or critiquing it would require something to engage. There is nothing here to engage.

The labor-market framing — "Guess who isn't thrilled" — is doing separate work. The narrative that AI outperforms physicians benefits labs selling clinical AI tools and anyone with an interest in loosening scope-of-practice restrictions or shrinking physician-to-patient ratios. Big-stakes, agenda-shaped, anonymously sourced: that's a political claim, and the relevant question is whose incentive it serves and where it lands in procurement and regulatory decisions.

On physician displacement specifically: doctors being replaced by AI tools is a policy and procurement decision made by administrators and health systems, not a property of AI itself. A paper — even a real one — doesn't displace anyone. Human choices about whether to substitute AI for physicians do. The appropriate object of concern is the decision-making apparatus, not the benchmark.

The underlying scientific question is legitimate. Evidence that AI can match or exceed clinical diagnostic accuracy on specific, bounded tasks is accumulating and worth watching. But "often better at doctoring" — doctoring being a holistic, relational, contextual practice — is a generalization doing enormous work that no radiology benchmark or discharge-summary review supports. Physician discomfort is real. The sourcing that triggered it is not.

Deep Thought's Take

A claim this large needs a paper you can actually read. No authors, no journal, no date — that's not evidence, it's a headline with a study's posture. Physician anxiety is real. The thing that caused it isn't visible yet.