Med Students Outsource Reasoning To Chatbots. Two Thirds Of Doctors Already Use One. The Never-Skilled Cannot Recover.

Simar Bajaj, a med student at Stanford, and Joseph Sakran, a trauma surgeon at Johns Hopkins, warn in The Guardian that medical students overusing AI risk "never-skilling" rather than just deskilling. Internal data shows roughly two thirds of US doctors already use OpenEvidence, a clinician-specific AI chatbot. The authors argue a doctor who forgot reasoning can relearn it, but one who never learned cannot.

This illustrates cognitive offloading, the gradual erosion of a skill you delegate to a machine before mastering it yourself. The principle is simple. Competence requires struggle. If you skip the struggle, the neural pathways never form. The same applies to anyone using AI to write, code, or think before they can do those things unaided. Use AI to refine your work, never to generate your understanding.

Simar Bajaj at Stanford University School of Medicine and Joseph Sakran at Johns Hopkins Medicine co-authored the Guardian essay. Their data shows approximately two thirds of US doctors use OpenEvidence, an AI chatbot designed for clinicians.