Artificial Intelligence-Induced Deskilling in Interventional Pulmonology: An International Cross-Sectional Survey on Risk Perception and Mitigation Strategies

Artificial intelligence (AI) is progressively reshaping interventional pulmonology (IP), yet its potential to erode procedural and cognitive competencies through AI-induced deskilling remains poorly characterized in this specialty. An international, observational, cross-sectional survey was conducted in May 2026 among 118 expert interventional pulmonologists from 10 different countries across 5 continents. Participants completed a structured questionnaire comprising five demographic items and 12 Likert-scale statements addressing deskilling risk perception and mitigation attitudes; percentage agreement was calculated for each item (scores 4–5). High perceived clinical value of AI was reported (87%), alongside substantial concern for procedural deskilling (73%) and upskilling inhibition (83%). Familiarity with automation bias was limited (38%), yet its clinical relevance was widely recognized after definition provision (81%)—a gap of 43 percentage points. Strong support emerged for AI-free training (84%), simulation-based training (86%), and longitudinal performance monitoring (78%). Concern for institutional fragility in the absence of AI was expressed by 74%, and governance frameworks, including minimum non-AI-assisted procedural volume requirements, were endorsed by 70%. Deskilling was identified as a high research priority by 89%. These findings indicate that AI-induced deskilling is perceived as a relevant and emerging risk by expert interventional pulmonologists internationally, even before the widespread clinical deployment of AI technologies. Although the extent to which these concerns will translate into measurable effects on procedural competence is currently uncertain, the results underscore the need for prospective research, educational initiatives, and appropriate governance frameworks to ensure the preservation of core procedural skills.