News

AI Assistance May Reduce Doctors' Skills When Technology is Unavailable, Study Suggests

New research indicates that endoscopists showed decreased polyp detection rates after exposure to artificial intelligence tools

Doctors who became accustomed to using artificial intelligence assistance during colonoscopy procedures showed reduced ability to detect abnormalities when the technology was unavailable, according to research published in The Lancet Gastroenterology & Hepatology.

The study, conducted across four endoscopy centers in Poland, found that physicians' adenoma detection rates decreased from 28.4% to 22.4% when performing standard colonoscopies after being exposed to AI-assisted tools—a 6-percentage-point decrease.

Researchers analyzed 1,443 patients who underwent colonoscopies between September 2021 and March 2022, comparing procedures conducted three months before and after the implementation of AI. The centers had introduced AI tools for polyp detection in late 2021 as part of the ACCEPT (Artificial Intelligence in Colonoscopy for Cancer Prevention) trial.

The AI systems used in the study placed visual markers around areas where abnormalities were detected, helping guide physicians' attention during procedures. However, the research suggests this assistance may have created an unintended dependency.

Broader Implications for Medical Practice
The findings contribute to growing concerns about over-reliance on automated systems in critical professions. Aviation industry experts have previously documented similar issues, where pilots became less proficient at manual flying after extensive exposure to autopilot systems.

The study's multivariable analysis identified three independent factors significantly associated with adenoma detection rates: exposure to AI (with an odds ratio of 0.69), patient sex, and patient age over 60 years.

Study Limitations and Future Research
The researchers acknowledged they had not anticipated this outcome and therefore did not collect data on the specific mechanisms behind the behavioral changes. The study was observational and retrospective, conducted at centers participating in the ACCEPT trial.

The research excluded patients with intensive anticoagulant use, pregnancy, or histories of colorectal resection or inflammatory bowel disease. All participants underwent diagnostic colonoscopies during the study period.

The findings raise questions about implementation protocols for AI assistance in medical settings, particularly regarding training methods that might prevent skill deterioration when technology is unavailable.

The study was funded by the European Commission and the Japan Society for the Promotion of Science.

About the Author

John K. Waters is the editor in chief of a number of Converge360.com sites, with a focus on high-end development, AI and future tech. He's been writing about cutting-edge technologies and culture of Silicon Valley for more than two decades, and he's written more than a dozen books. He also co-scripted the documentary film Silicon Valley: A 100 Year Renaissance, which aired on PBS.  He can be reached at [email protected].

Must Read Articles

Welcome to MedCloudInsider.com, the new site for healthcare IT Pros looking for insights on cloud and other cutting-edge IT tech.
Sign up now for our newsletter and don’t miss out! Sign Up Today