What happened
Physicians using an AI tool for colonoscopy analysis experienced a decline in their ability to detect pre-cancerous adenomas when the AI was unavailable. A Polish study, published in The Lancet Gastroenterology and Hepatology, found that after using the AI, specialists' adenoma detection rate dropped from 28.4% to 22.4% during colonoscopies performed without AI assistance. This follows a survey indicating 70% of nurses and 77% of physicians worry about skill loss from AI over-reliance. Separately, Anthropic researchers conducted a trial with 52 software engineers, half using an AI assistant for a basic coding task, to investigate skill loss in computer science.
Why it matters
Reliance on AI tools risks atrophy of critical human expertise, directly impacting patient outcomes and operational resilience. For medical specialists, this means a measurable reduction in diagnostic accuracy without AI support, increasing patient risk. Procurement teams must recognise that integrating AI can introduce new dependencies, potentially degrading core competencies if human skill maintenance is not prioritised. This follows prior research indicating AI chatbots misdiagnose patients, underscoring the need for robust human oversight.
