The more professionals rely on artificial intelligence, the more their own skills erode — that is the stark conclusion of mounting evidence surveyed by Nature magazine in a feature published June 18, 2026.

A study of Polish endoscopy specialists in The Lancet Gastroenterology & Hepatology found that after they began using an AI tool that flags precancerous adenomas in real time, their unaided adenoma detection rate dropped from 28.4% to 22.4% — a statistically significant decline. The physicians, all of whom had performed at least 2,000 colonoscopies in their careers, appeared to lose motivation and focus when working without the AI.

"There is no established solution against deskilling right now," said co-author Yuichi Mori, a physician-researcher at the University of Oslo. "It should be a very hot research topic in the next decade."

In computer science, a randomized controlled trial by AI firm Anthropic tested 52 software engineers on a basic coding task. Half were prompted to use an AI assistant. The result: those who used AI produced worse code than those who relied solely on their own knowledge, web searches, and documentation.

A 2026 survey of U.S. healthcare workers found that 70% of nurses and 77% of physicians are already worried about losing their skills because of over-reliance on AI systems.

Researchers say the phenomenon — known as "deskilling" — has been observed before in other industries. A landmark 2018 study by Rinta-Kahila and colleagues found that tax accountants who relied on software automation lost the ability to spot errors the system missed.

"Just being aware that this phenomenon exists hopefully provokes some self-reflection about which skills people want to maintain and which they're willing to outsource," said Kevin Crowston, an information scientist at Syracuse University.

The findings arrive as AI adoption accelerates across medicine, law, software, and finance. The question of how to preserve human expertise in an age of increasing automation remains unanswered.

Sources