A new study in The Lancet Gastroenterology & Hepatology warns that frequent use of artificial intelligence in medical procedures may gradually weaken doctors’ core clinical skills. Researchers in Poland found that experienced physicians who regularly relied on AI during colonoscopy screenings showed a significant drop in their ability to detect pre-cancerous growths when the AI system was temporarily removed. Their adenoma detection rate decreased from 28.4 percent to 22.4 percent, suggesting that routine dependence on AI can lead to reduced focus and confidence, a phenomenon known as automation bias.
Despite this caution, experts emphasize that AI has already delivered major benefits to modern medicine. AI imaging tools can spot cancers earlier than traditional methods, AI-assisted radiology can reduce errors, and predictive models can help identify patients at high risk of conditions such as sepsis or heart failure. In many hospitals, AI-supported triage systems have improved emergency room efficiency and helped clinicians make faster decisions in high-pressure situations. These successes demonstrate the value of artificial intelligence as a supportive partner that strengthens patient care.
The challenge, researchers say, is finding the right balance. Doctors must continue to practice and refine their skills without technological assistance so they remain fully capable when AI is unavailable or when human judgment is essential. The study encourages ongoing evaluation of how AI affects clinical performance over time, along with training methods that preserve human expertise.
AI is becoming an increasingly powerful ally in healthcare, but the study highlights the importance of maintaining strong medical fundamentals. The goal is to integrate artificial intelligence in ways that enhance the abilities of clinicians while ensuring that human skill, attention, and decision-making remain at the center of patient care.