Overview
- A multicentre observational study in four Polish centres found non-AI colonoscopy adenoma detection rates fell from 28.4% to 22.4% after three months of routine AI assistance.
- Overall ADR rose to 25.3% when AI was active, but AI-assisted procedures did not significantly outperform pre-AI detection baselines.
- Study authors and commentators suggest overreliance on AI may weaken visual search habits, gaze patterns and diagnostic confidence.
- Researchers acknowledge the study’s observational design, single-system use and experienced clinician cohort limit causal claims and generalisability.
- Experts are calling for professional-society guidelines, regular non-AI practice sessions and randomized crossover trials to safeguard endoscopist skills and patient outcomes