AI-Assisted Colonoscopies Linked to Drop in Manual Adenoma Detection: Lancet Study

As Artificial Intelligence (AI) becomes increasingly integrated into medical practice, a new study published in The Lancet Gastroenterology and Hepatology has raised concerns about its potential impact on clinician performance. The research suggests that doctors who routinely perform AI-assisted colonoscopies may lose their ability to independently detect precancerous growths—known as adenomas—in the colon.

Colonoscopy, performed by endoscopists, is a key procedure used to detect and remove adenomas, thereby preventing bowel cancer. While previous studies have shown that AI can enhance adenoma detection rates, this study questions whether long-term reliance on the technology may erode essential clinical skills.

“To our knowledge, this is the first study to suggest a negative impact of regular AI use on healthcare professionals' ability to complete a patient-relevant task in medicine of any kind. Our results are concerning, given that the adoption of AI in medicine is rapidly spreading,” said Dr. Marcin Romańczyk from the Academy of Silesia, Poland.

The study was conducted across four colonoscopy centres in Poland between September 2021 and March 2022. It involved 1,443 non-AI-assisted colonoscopies—795 performed before the introduction of AI, and 648 after its implementation. All procedures were carried out by 19 seasoned endoscopists, each with experience in over 2,000 colonoscopies.

The findings showed a significant drop in adenoma detection rates in procedures done without AI following its routine adoption. Detection rates fell from 28.4 per cent (226/795) before AI exposure to 22.4 per cent (145/648) afterwards—reflecting a 20 per cent relative and 6 per cent absolute decrease. In comparison, AI-assisted colonoscopies showed a 25.3 per cent (186/734) detection rate.

The team also acknowledged some limitations, including that the observational nature of the study means that factors other than the implementation of AI use may have influenced the findings.

Romańczyk called for more research into the impact of AI on health professionals’ skills across different medical fields.