Priya Dave

Navigating the Ethical Landscape of AI in Radiology: Challenges and Frameworks for Responsible Integration

Priya Dave

Mayo Clinic

dave.priya2@mayo.edu

Artificial intelligence (AI) has made significant strides in the field of radiology, with radiology emerging as the frontrunner in AI tool development and FDA approvals. As of 2024, radiology accounts for more than three-quarters of AI-enabled devices authorized by the FDA. AI in radiology has shown great potential to enhance diagnostic accuracy, improve efficiency, and advance patient care. However, the integration of AI technologies in radiological practice raises significant ethical concerns that must be carefully addressed.

1) Automation Bias: One of the primary ethical concerns in AI-assisted radiology is automation bias. Studies have shown that radiologists, regardless of their experience level, are prone to automation bias when supported by AI systems. This bias can lead to over-reliance on AI-generated results, potentially compromising diagnostic accuracy and patient safety. The risk of automation bias is heightened in situations of fatigue or limited radiological workforce.

2) Data Ethics: The development and implementation of AI in radiology rely heavily on large datasets. Ensuring the ethical use of patient data, including informed consent, privacy protection, and data ownership, is crucial.

3) Bias: AI systems may inadvertently perpetuate or amplify existing biases, leading to unfair treatment of certain patient groups.  This can result in misdiagnoses, inappropriate care, or reduced access to healthcare resources for underrepresented populations.

4) Transparency: The ""black box"" nature of some AI algorithms poses challenges in understanding and explaining AI-driven decisions. Ensuring transparency in AI systems is crucial for maintaining trust, facilitating informed decision-making, and establishing clear lines of responsibility and accountability

5) Responsibility: As AI systems become more autonomous, questions arise regarding responsibility and liability for AI-assisted diagnoses. Defining the roles and responsibilities of human radiologists and AI systems is essential for ethical practice and patient safety.

A framework for understanding AI in radiology can serve as a model for AI integration across other fields. Case studies illustrating each ethical challenge and its resolution will provide valuable insights and practical guidance for healthcare professionals navigating this rapidly evolving landscape.

BIOGRAPHY

Priya Dave is a radiology resident physician at Mayo Clinic. She earned her medical degree from the Icahn School of Medicine at Mount Sinai and master's in bioethics from Harvard Medical School, where her master's thesis focused on moral distress in radiology. Priya’s academic interests center on the intersection of bioethics, humanities, and radiology. Her ethics work has appeared in the American Journal of Radiology. Priya has furthered her interests in bioethics through studies at the Oxford University Ethox Center. Currently, she serves as a Board Member of the American Osler Society and co-leads the American Association of Radiologists Humanities Committee. In addition to her clinical and academic work, Dr. Dave has an interest in the history of medicine.