AI Summary & Key Takeaways
AI InsightResults: The overall mean DISCERN score indicated moderate information quality (39.2±9.7).
📌 3 Key Takeaways
Abstract
Aims: Artificial Intelligence (AI) has rapidly gained prominence in radiology, leading to widespread public discourse on whether AI can replace radiologists. YouTube has become a major source shaping this perception; however, the reliability, transparency, and ethical quality of such content remain uncertain. This study aimed to evaluate the information quality of YouTube videos addressing the potential replacement of radiologists by AI. Methods: This cross-sectional observational study analyzed publicly available YouTube videos published between 2017 and 2025. A total of 183 eligible videos were evaluated using validated assessment tools, including the DISCERN instrument, Journal of the American Medical Association (JAMA) benchmark criteria, and health on the net code of conduct (HONcode) principles. Video characteristics, uploader type, content category, and engagement metrics were recorded. Two independent reviewers assessed all videos, and inter-rater reliability was calculated using Cohen’s kappa coefficient. Results: The overall mean DISCERN score indicated moderate information quality (39.2±9.7). Videos uploaded by academic institutions demonstrated significantly higher DISCERN scores than those uploaded by individual users (46.5±7.9 vs. 36.0±9.1; p<0.001). None of the videos fulfilled all four JAMA benchmark criteria, while only 39.3% of the videos complied with at least six HONcode principles. Engagement metrics showed weak but statistically significant associations with information quality: view count (r=0.18, p=0.012) and video duration (r=0.19, p=0.009) were weakly positively correlated with DISCERN scores, whereas the like-to-view ratio demonstrated a weak negative correlation (r=-0.16, p=0.033). Inter-rater reliability for DISCERN scoring was near perfect, with a Cohen’s kappa coefficient of 0.87. Conclusion: YouTube videos discussing whether AI can replace radiologists exhibit moderate information quality with substantial deficiencies in transparency and ethical standards. Academic institution-produced content is more reliable than non-academic sources, while popularity metrics do not reliably indicate information quality.