فهرست نشریات علمی دسترسی آزاد ایران

Background and Objectives: Advances in machine learning and deep learning have accelerated artificial intelligence (AI) development in radiology, yet clinical adoption remains limited due to trust, accountability, and acceptance concerns. This study assessed Iranian medical residents’ attitudes toward AI in radiology.    Methods: In this cross-sectional study, 112 specialty residents from four teaching hospitals affiliated with Ahvaz Jundishapur University of Medical Sciences voluntarily completed a validated 39-item questionnaire covering five domains: Distrust and Responsibility, Procedural Knowledge, Human Interaction, Efficiency, and Being Informed. Items were rated on a 5-point Likert scale, with reverse scoring applied where appropriate (higher scores indicating more negative attitudes). Descriptive statistics summarized responses, group comparisons used parametric or non-parametric tests, and associations were analyzed with Spearman correlation (P < 0.05).   Results: Participants’ mean age was 31.84 ± 5.20 years; 56.3% were female, 67.0% were in non-surgical specialties. Overall, 92.9% had moderate attitudes toward AI; none were classified as positive. Distrust and Responsibility had the highest mean score (48.99 ± 4.90), reflecting concerns about accountability and reliability. Attitude scores did not differ by gender or specialty. Age showed a weak but significant positive correlation with total attitude score (ρ = 0.272, P = 0.004) and with subscales of Distrust and Responsibility and Human Interaction.    Conclusion: Medical residents exhibited generally cautious and moderate attitudes toward AI, mainly due to trust and professional responsibility concerns. Single-center design and voluntary participation may introduce selection bias. Targeted education and clearer ethical and legal frameworks could address these concerns and support responsible AI integration.