Study Finds That AI Use Degrades Doctors’ Competence

Category :

AI

Posted On :

Share This :

 

The first clinical evidence that regular use of artificial intelligence can lead to healthcare professionals losing critical skills was presented in a groundbreaking study published in The Lancet Gastroenterology and Hepatology. This raises serious concerns about the widespread adoption of AI in medical practice.

 

Dr. Marcin Romanczyk of the Academy of Silesia in Poland and associates conducted the study, which looked at 1,442 colonoscopies carried out by skilled endoscopists at four Polish hospitals between September 2021 and March 2022. When doctors’ performance was compared before and after AI-assisted polyp detection technologies were introduced, researchers noticed a worrying trend they called “deskilling.”

 

AI Has A Major Effect On Diagnostic Abilities

The capacity of endoscopists to identify precancerous adenomas during routine, non-AI assisted colonoscopies decreased from 28.4% to 22.4% after three months of consistent AI use, a 20% relative and 6% absolute decline, according to the study. This was the first recorded instance of AI resulting in a quantifiable loss of clinical skills among healthcare workers, according to STAT News.

 

“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,” stated Dr. Romanczyk. The findings are particularly troubling given that the endoscopists evaluated were highly experienced, having each performed over 2,000 colonoscopies.

 

 

Professional Responses Emphasize Wider Consequences

The study has caused medical professionals to react in different ways. The study “provides the first real-world clinical evidence for the phenomenon of deskilling, potentially affecting patient-related outcomes,” according to Dr. Omer Ahmad, a consultant gastroenterologist at University College London. Some experts, however, have advised using caution when interpreting the findings.

 

After implementing AI, the number of colonoscopies almost doubled, according to Professor Venet Osmani of Queen Mary University of London. This suggests that the performance decline may be due to increasing workload rather than AI alone. He questioned whether therapists with decades of expertise, in particular, could endure true skill loss in just three months.

 

 

Concerns Over Healthcare Automation Are Growing

The study comes as worries about how AI will affect the abilities of medical professionals in many professions are mounting. Similar “automation bias” effects, when experts become unduly dependent on AI systems, have been demonstrated in various sectors by recent study. Frequent usage of AI assistants that simulate cognitive processes is expected to result in more skill erosion than traditional automation systems, according to a 2024 study published in Clinical Psychological Science.

 

“Health professionals who become accustomed to using AI support will perform more poorly than they originally did if the AI support becomes suddenly unavailable, for example due to cyber-attacks or compromised IT systems,” cautioned Dr. Catherine Menon of the University of Hertfordshire.

 

The ramifications go beyond certain professionals. The study found that as people lose their ability to function without digital support, deskilling weakens and weakens healthcare systems. Serious repercussions could result from this vulnerability in the event of system failures or anomalous medical presentations that defy AI training parameters.

 

The release of the report comes as major medical organizations have issued conflicting recommendations regarding the use of AI in colonoscopies. While BMJ Rapid Recommendations advises against routine AI usage in general colonoscopy screening, the European Society of Gastrointestinal Endoscopy advocates its deployment based on patient preference assumptions. As healthcare institutions around the world increasingly use AI technologies, these contradictory suggestions illustrate the difficult balance between the hazards and advantages of AI.