Artificial intelligence (AI) is widely recognized for its ability to improve diagnosis and disease management. However, a new study published in Lancet Gastroenterology and Hepatology has raised an important concern: the risk of clinician deskilling due to constant reliance on AI-based decision-support systems.
The Study Setup
Researchers conducted the study across four colonoscopy centers in Poland and introduced AI tools for detecting polyps and abnormal colon cells at the end of 2021. They aimed to evaluate whether prolonged exposure to AI assistance influenced doctors’ diagnostic accuracy.
Alarming Drop in Detection Rates
The findings were striking. The average adenoma detection rate (ADR)—the ability to identify non-cancerous growths that could develop into cancer—fell significantly in non-AI assisted colonoscopies after prolonged AI use. Detection dropped from 28% before AI exposure to 22% afterward, amounting to a 20% relative and 6% absolute reduction in ADR.
Researchers Express Concern
“To our knowledge, this is the first study to suggest a negative impact of regular AI use on healthcare professionals’ ability to perform a patient-relevant task in medicine of any kind,” said Dr. Marcin Romarnczyk of the Academy of Silesia, Poland. “Our results are concerning given the rapid adoption of AI in medicine. We urgently need more research into how AI affects clinicians’ skills across different fields and which factors contribute to this problem.”
Yuichi Mori of the University of Oslo added that the findings raise questions about earlier randomized controlled trials (RCTs) showing AI-assisted colonoscopy improved ADR. “It could be that non-AI assisted colonoscopies in these RCTs were not the same as standard procedures, as endoscopists may have been negatively affected by continuous AI exposure,” he explained.
The study received funding from the European Commission, the Japan Society for the Promotion of Science, and the Italian Association for Cancer Research.
Why Deskilling May Occur
According to the researchers, continuous exposure to AI and decision-support systems can trigger a natural human tendency to over-rely on external recommendations. This reliance may lead clinicians to become less motivated, less attentive, and less responsible in exercising their own diagnostic skills.
Balancing Risks and Benefits
While the study highlights potential risks, some experts urge a balanced view. Dr. Vidur Mahajan, Founder and CEO of CARPL.AI, argued that the focus should be on enabling average doctors to perform at the level of the best in the world, rather than fearing deskilling.
“Technology is an inevitable part of our lives, and we must embrace its advantages by democratizing access to it. Imagine a world without Google Maps—would you trust a driver who refuses to use it?” he said.
The Road Ahead
As reported by TOI, the study underscores the importance of examining not just the benefits but also the unintended consequences of AI in medicine. As AI adoption accelerates, ongoing research, training, and thoughtful integration will be essential to ensure that technology enhances—rather than diminishes—clinicians’ skills and patient outcomes.