Why look at doctors' vigilance?
A colonoscopy uses a small camera to look inside the large bowel. The doctor looks for adenomas, polyps (small growths) that can turn into cancer over the years. Removing them early helps prevent bowel cancer.
AI tools flag suspicious areas on screen during the test, and they help find more adenomas. But one question remained open: after relying on this help, do doctors keep the same sharp eye when it is switched off?
What did the researchers do?
The team studied four endoscopy centers in Poland that were part of a large trial of AI in colonoscopy. At the end of 2021 these centers received AI tools, and from then on tests were done with or without AI depending on the date.
The researchers compared colonoscopies done without AI in the 3 months before AI arrived and in the 3 months after. They looked at the adenoma detection rate, meaning the share of tests in which at least one adenoma is found. It is a standard measure of colonoscopy quality.
What were the results?
In total, 1,443 people had a colonoscopy without AI: 795 before and 648 after AI was introduced. The adenoma detection rate fell from 28.4% to 22.4%, a drop of 6 points.
After taking other factors into account, exposure to AI was still linked to lower detection, as were patients' sex and age. The authors see this as a possible negative effect of getting used to AI on doctors' behavior.
Key figures
- 1,443colonoscopies done without AI, before and after AI was introduced
- 28.4% then 22.4%adenoma detection rate without AI, before and then after getting used to AI
- 4endoscopy centers in Poland
Getting ready for your colonoscopy with Mima
This study does not change the value of having a colonoscopy when you are offered one. What it does show is that the quality of the test matters, and that you can talk about it with the team doing it.
With Mima, you can keep track of your appointments and results, and prepare your questions in advance. For example: "Do you use an AI tool during the test, and how do you keep an eye on the quality of your colonoscopies?"
What could this mean for patients?
This study does not say that colonoscopy is less safe, or that AI is a bad idea. It looks at tests done without AI, once doctors had become used to having it. It also does not measure whether any cancers were missed.
Above all, it raises a question for how care is organized: how can doctors keep their skills sharp when they work alongside AI? Ongoing training and monitoring the quality of tests remain essential.
What are the limits?
This is an observational before-and-after study covering a short period of 3 months on each side. Other changes at the same time may have played a part, and the study cannot prove that AI caused the drop.
It only covers four centers in one country. Further, longer studies in other hospitals are needed to see whether the effect holds and how to prevent it.
Frequently asked questions
Should I avoid an AI-assisted colonoscopy?
No. The study did not look at the quality of tests done with AI, and colonoscopy remains a useful test to help prevent bowel cancer.
What is an adenoma?
It is a bowel polyp, a small growth that can turn into cancer over the years. Removing it during a colonoscopy helps prevent that cancer.
Does AI make doctors worse?
The study shows lower detection without AI after doctors got used to it, from 28.4% to 22.4%. As it is observational, it does not prove that AI is the cause.
What does the adenoma detection rate measure?
The share of colonoscopies in which at least one adenoma is found. It is a common measure of test quality.
The original study
Budzyń K, Romańczyk M, Kitala D, et al. “Endoscopist deskilling risk after exposure to artificial intelligence in colonoscopy: a multicentre, observational study” The Lancet Gastroenterology & Hepatology, August 12, 2025. DOI : 10.1016/S2468-1253(25)00133-5
Read the full study in The Lancet Gastroenterology & HepatologyThis article summarizes a scientific study and is not medical advice.