Responsible Integration of Artificial Intelligence In Radiation Medicine: A Pan-Canadian Framework for Oversight, Education, and Collaboration
Abstract
Purpose
Artificial intelligence (AI) is increasingly embedded across radiation medicine, with applications in clinical decision support, workflow efficiency, personalization of care, and quality assurance. Despite rapid technical advancement, there remains a critical gap in guidance on how AI should be responsibly integrated into clinical practice. This work presents a pan-Canadian, interprofessional framework for preparing radiation medicine for an AI-enabled future, emphasizing governance, education, collaboration, and patient-centred values as essential enablers of safe and effective implementation.
Methods
The work synthesizes current literature, professional guidance, and Canadian health system context to examine emerging AI applications alongside ethical, professional, and operational considerations relevant to radiation medicine. Using a consensus-based approach, it evaluates system-level requirements for AI integration, including oversight structures, data governance, workforce preparation, and collaborative implementation models, with particular attention to the role of medical physicists in commissioning, validation, and quality assurance.
Results
Three interconnected domains essential to responsible AI integration are identified. First, AI must be implemented within robust governance and oversight frameworks, supported by human supervision and AI-ready data standards, to ensure transparency, reproducibility, and clinical reliability. Second, comprehensive AI literacy across entry-to-practice education and continuing professional development is required to support evolving professional roles, including physics leadership in AI evaluation, commissioning, workflow integration, and ongoing performance monitoring. Third, coordinated collaboration across institutions, professions, and jurisdictions enables shared validation efforts, early adoption, and alignment with national and international informatics standards. Across all domains, patient trust, accountability, transparency, and equity are identified as foundational requirements, with increasing emphasis on patient participation in AI governance and evaluation.
Conclusion
AI will significantly shape the future of radiation medicine, but its impact will depend on how deliberately it is implemented. Aligning technological innovation with rigorous governance, education, collaboration, and patient-centred values provides a practical pathway for responsible, sustainable, and high-quality AI-enabled care.