A multisociety collaboration among the American Association of Physicists in Medicine (AAPM), American College of Radiology (ACR), Radiological Society of North America (RSNA), and Society for Imaging Informatics in Medicine (SIIM) developed a comprehensive educational framework to address the growing complexity of AI integration in medical imaging. Published in Medical Physics, Radiology: Artificial Intelligence, and the Journal of Imaging Informatics in Medicine, the syllabus responds to challenges posed by the complexity of machine learning models, substantial data requirements, and the potential for bias and error. Nearly 30 experts from the four societies collaborated on this initiative to establish standardized educational competencies.
The syllabus outlines competencies for four distinct personas: clinical users of AI systems, purchasers of AI systems, clinical collaborators who provide expertise during AI development, and developers of AI systems. Intentionally designed as a syllabus rather than a curriculum, the framework provides flexibility for educators to adapt content while maintaining consistent instruction on AI fundamentals, clinical integration, regulatory requirements, and ethical considerations. Key competencies include establishing policies for implementation, ongoing monitoring, and continual evaluation of algorithms to detect potential degradation in accuracy or adverse effects, with emphasis on governance committees that include diverse stakeholders throughout the model acquisition and deployment lifecycle.