A new review of research published in Nature warns that growing reliance on artificial intelligence may cause “deskilling,” meaning professionals gradually lose abilities they built over years by handing more cognitive tasks to AI. The article says the technology can improve short-term performance, but may also weaken independent work over time, from medicine to software development.
The concern is already visible in U.S. healthcare. A survey found that 70% of nurses and 77% of doctors fear heavy dependence on AI will hurt their professional capabilities. One study, published in The Lancet Gastroenterology & Hepatology, followed expert colonoscopy doctors after they began using a real-time AI system to spot precancerous lesions. When the AI was unavailable, their lesion-detection rate fell by about 21%.
The researchers said prolonged use may reduce doctors’ concentration, motivation and responsibility when making decisions without AI support. Dr. Yoichi Mori of the University of Oslo, one of the paper’s authors, said more research is needed to confirm the scale of the problem, but warned users should realize they may lose some professional skills. He added that there is still no proven way to prevent deskilling and expects the issue to become a major research field in the next decade.
A separate Anthropic study involving 52 software engineers found a similar pattern. Half used an AI assistant during a coding task and were then tested on what they had learned. Those who used AI scored an average of 50%, compared with 67% for those who did not. The biggest gap was in questions requiring debugging and understanding code, suggesting participants could finish the task but understood the underlying principles less well. The researchers said the risk is especially serious for students and junior developers, and recommended continuing to practice skills without AI, learning how the systems work, and not accepting their answers uncritically.
