Evidence of AI-driven deskilling emerges in studies
Nature reports early studies finding routine AI assistance can measurably degrade professionals' core skills, notably in medicine and software engineering. A Lancet Gastroenterology and Hepatology study published last October found clinicians were worse at detecting precancerous growths when AI support was removed. Study authors say continuous AI reliance made clinicians less motivated, less focused, and less responsible when making independent decisions, and separate experiments found heavy language-model use reduced problem-solving performance on programming tests. These findings matter because they risk patient safety, weaken training pipelines, and create brittle workforces that may fail if AI systems are unavailable or wrong, prompting calls to preserve human expertise and redesign training and oversight.
