Medical experts have raised concerns that the growing use of artificial intelligence (AI) in colonoscopy procedures may be undermining physicians’ natural abilities to detect tumors without technological assistance. A comprehensive study published in the journal Lancet Gastroenterology and Hepatology investigated how the introduction of AI tools for polyp detection affected the performance of endoscopists. The research, conducted at four Polish endoscopy centers, followed patients before and after the implementation of AI in late 2021.
A total of 1,443 patients underwent colonoscopies between September 2021 and March 2022, with some procedures conducted with AI assistance and others without. The tumor detection rate, measured by the adenoma detection rate (ADR), declined from 28.4% to 22.4%—a 6% reduction—after AI tools were introduced. The study highlighted that AI exposure became an independent factor associated with the decline in ADR, suggesting a negative impact on endoscopists’ behaviors and skills.
Harvey Castro, an emergency physician and specialist in AI and healthcare, described the study as an important piece of work. He noted that while AI can assist in detecting cancer, it may also reduce the doctors’ ability to identify tumors when the technology is not available. Castro emphasized that even small changes in detection rates can significantly affect patient survival, particularly for younger adults experiencing rising colorectal cancer diagnoses. He called for a balanced approach, advocating for the wise integration of AI to enhance, rather than replace, the expertise of medical professionals.
The findings prompted discussions among medical professionals about the future of AI in healthcare. Critics argue that over-reliance on technology without proper oversight could lead to long-term skill degradation among practitioners. Meanwhile, proponents of AI in medicine maintain that these tools can improve detection rates and save lives, provided they are used in conjunction with human expertise. The study serves as a cautionary tale about the need for training, oversight, and intentional design when incorporating AI into clinical practice to ensure the safety and effectiveness of patient care.