Wolfgang Miesbach: The Risk of Deskilling in Modern AI Assisted Medicine
Wolfgang Miesbach, Professor of Medicine at Frankfurt University Hospital, shared a post on LinkedIn about a recent article by Yuhe Ke et al., published in Nature Medicine, adding։
“Loosing your skills by AI?
This Perspective in Nature Medicine shows how AI can undermine clinical competence: students risk ‘never‑skilling’ (they never build independent diagnostic schemas), while experienced clinicians risk deskilling and mis‑skilling (eroding existing skills or internalising AI errors as ‘truth’).
The problem is not AI itself, but overuse of ‘answer‑delivery mode’ that bypasses the hard cognitive work needed for robust reasoning, calibration and metacognitive judgment.
The authors propose a 3‑step approach:
- AI‑free reasoning zones (history, exam, DDx, management) for trainees andperiodic “AI‑off” drills for seniors,
- guided learning‑mode AI with Socratic questioning and built‑in errors to train verification/override ,
- integrated human-AI collaboration where clinicians contribute context, values and responsibility instead of just relaying outputs .
The question for education, leadership and regulators: how do we ensure that, in an AI‑rich world, every clinician can still fly manually when the ‘autopilot’ fails?”
Title: AI-induced never-skilling in medical education
Authors: Yuhe Ke, Liyuan Jin, Jasmine Chiat Ling Ong, Arun J. Thirunavukarasu, Josip Car, Carol Y. Cheung, Yih Chung Tham, Daniel Shu Wei Ting, Marcus Eng Hock Ong, Scott Compton, Aditee Narayan, Pearse A. Keane, Tien Yin Wong, David W. Bates, Patrick Tan, Nan Liu
Other posts featuring Wolfgang Miesbach on Hemostasis Today.
-
May 29, 2026, 12:20Pall T. Onundarson: Is Anonymous Peer Review Always Objective
-
May 29, 2026, 11:59Jolene Ng: New Review Explores the Pathophysiology of Pre-Eclampsia
-
May 29, 2026, 11:44Lynne Baird: Why Expanding Bleed Control Kits Across Public Transport Matters
-
May 29, 2026, 11:30Rumbidzai Mushambi: When ‘Just Hormones’ Delays Real Diagnosis
-
May 29, 2026, 09:53Yassin Taher: Recognizing Clonal Eosinophilic Disorders in Pediatric Hypereosinophilic Syndrome
-
May 29, 2026, 09:52Sergio Emanuel Kaiser: Ruvonoflast and a New Anti-Inflammatory Strategy for Residual Cardiovascular Risk
-
May 29, 2026, 09:50Maria Symiakaki: Advancing Cardiovascular Prevention at the 2026 Congress of the European Atherosclerosis Society
-
May 29, 2026, 09:37Katie Baca-Motes: What Can Menstrual Cycle Physiology Tell Us About Health Beyond Reproduction?
-
May 29, 2026, 09:35Torben Boelk: Bringing Plasma Donation to the Heart of Leipzig