Annals of Atrophy: Doctors struggle with diagnoses after becoming AI dependent

A new study published in The Lancet Gastroenterology & Hepatology reveals that doctors who rely on artificial intelligence for medical procedures may be experiencing “deskilling”—a gradual loss of diagnostic abilities when the technology isn’t available. Researchers found that experienced endoscopists (doctors who perform colonoscopies) became significantly less effective at detecting precancerous polyps during colonoscopies after becoming accustomed to AI assistance, with detection rates dropping from 28.4% to 22.4% when the technology was removed.

What you should know: The study tracked experienced physicians across four endoscopy centers in Poland who alternately performed colonoscopies with and without AI assistance.

• All participants were highly experienced clinicians, ruling out fatigue or inadequate training as factors in the performance decline.
• AI was identified as the only modifiable factor that could explain the significant drop in polyp detection rates.
• The findings surprised even the researchers, who weren’t expecting such a substantial difference in diagnostic quality.

Why this matters: Colonoscopies are a primary method for preventing colorectal cancer by identifying and removing adenomas (precancerous growths), making accurate detection crucial for patient outcomes.

• AI systems designed to improve polyp detection have shown promise in increasing diagnostic rates in real time.
• However, the study suggests that over-reliance on these tools might create a dependency that compromises doctors’ natural diagnostic skills.
• This raises concerns about the long-term implications of rapid AI adoption in healthcare settings.

What the researchers are saying: Dr. Marcin Romańczyk of the Academy of Silesia, one of the lead study authors, drew parallels to everyday technology dependence.

• “We have to admit that we were not expecting to find out such a big difference in quality parameters of colonoscopy,” Romańczyk told Newsweek.
• “I don’t personally feel confident anymore traveling on my own without any software assistance to an exact spot in another city, even within my region, because I’ve got used to using online map systems.”
• “On one side, AI usage is becoming standard of care. On the other hand, we are aware that we can not correctly state how our behavior has been modified by its usage.”

Expert commentary: University College London’s Dr. Omer Ahmad cautioned against unchecked enthusiasm for AI adoption in healthcare.

• “These findings temper the current enthusiasm for rapid adoption of AI-based technologies […] and highlight the importance of carefully considering possible unintended clinical consequences,” Ahmad wrote in a commentary accompanying the study.

What’s next: The study authors are calling for expanded research into AI’s impact on medical professionals across different specialties.

• They emphasize the need for solutions to prevent skill decay, including revised training protocols or limits on continuous AI use.
• “In our opinion, usage of AI is inevitable, and we should do our best to be prepared as well as possible to achieve the best outcomes from AI assistance,” Romańczyk concluded.