A multicenter study published in The Lancet Gastroenterology and Hepatology found that clinicians who used artificial intelligence during colonoscopies later performed worse when the technology was removed. Researchers working across four endoscopy centers in Poland observed a decline in polyp detection: the adenoma detection rate fell from 28.4% with assistance to 22.4% when clinicians performed procedures without software support. The difference surprised the authors and raised the possibility of a deskilling effect linked to routine reliance on decision support tools.
The trial focused on experienced endoscopists who had incorporated AI into their workflow to help spot precancerous polyps in real time. According to the paper and correspondence with Newsweek, the drop in performance was not attributable to fatigue or lack of training; all participants were highly experienced and AI was the only modifiable factor. Dr. Marcin Romańczyk of the Academy of Silesia, a lead author, said, ´We have to admit that we were not expecting to find out such a big difference in quality parameters of colonoscopy.´ He likened technology dependence to everyday examples, noting that people stop navigating as confidently when they always use electronic maps.
The study balances two realities: AI systems have shown promise in improving diagnostic rates during live procedures, and rapid adoption risks unintended consequences. Authors called for broader research into whether similar effects appear in other specialties, and for practical responses such as revised training protocols or limits on continuous AI use to prevent skill decay. In an accompanying commentary, University College London’s Dr. Omer Ahmad warned that the findings ´temper the current enthusiasm for rapid adoption of AI-based technologies´ and underline the need to consider unintended clinical consequences. The paper´s authors concluded that integration of artificial intelligence into care is likely inevitable, but preparation, monitoring and safeguards will be essential to retain clinician expertise while reaping the benefits of new tools.
