Abstract / Summary
Abstract Introduction: The integration of artificial intelligence (AI) into radiology promises transformative advancements while simultaneously raising complex ethical dilemmas. This study examines the ethical perspectives of radiologists in Türkiye through the Q methodology, which systematically maps practitioners’ subjective viewpoints. Methods: This descriptive cross-sectional study employed Q methodology. Data were collected between November 17, 2023, and April 4, 2024 from radiologists practicing in Türkiye who actively used or supervised AI applications in medical imaging, with snowball sampling using a digital Q-sort instrument developed by the researchers and available at https://ethicsqsort.netlify.app. Participants ranked 36 statements on a 7-point Likert scale ranging from -3 (strongly disagree) to +3 (strongly agree). Data were analyzed using person-centered factor analysis with KADE software. Results: A total of 41 radiologists were included in the study. Three distinct ethical perspectives emerged regarding AI use in radiology: a transparency–patient autonomy perspective, an efficiency–pragmatism perspective, and a reliability–justice perspective. These viewpoints diverged most clearly on the necessity of informing patients about AI involvement, expectations for explainability and attitudes toward data privacy. Despite these differences, strong convergence was observed across all factors. Participants agreed that AI should function as a supportive tool, collaboration with system designers is essential for reliable use, and occasional errors do not justify discontinuing the technology. They also anticipated that radiologists with AI skills will be preferentially employed in the future. Conclusion: Despite differing ethical orientations radiologists with AI experience shared core principles that support responsible clinical AI integration. These findings highlight the need for national guidelines in Türkiye to clarify transparency, data governance, and clinician oversight. Strengthening multidisciplinary governance, improving AI literacy, and promoting clinician–developer collaboration will be essential for trustworthy and equitable adoption of AI in radiology.