Developing Competency in AI Ethics: A training framework

AI ethics requires practical, step-by-step training: The framework breaks down abstract ethical principles into four concrete tasks—identifying concerns, mapping requirements, selecting measures, and recognizing trade-offs—making ethics actionable for researchers, engineers, policymakers, and students rather than leaving it as theoretical guidance.

Ethics tensions are inevitable and need explicit reasoning: The framework acknowledges that ethical requirements often conflict with each other, and rather than pretending perfect solutions exist, it trains practitioners to explicitly identify these tensions and reason through trade-offs, reflecting real-world complexity in AI development.

Multiple stakeholders need ethics competency: The training is designed for diverse audiences including Research Ethics Committee members, policymakers, and students, recognizing that responsible AI requires ethics literacy across different roles and organizational levels, not just among technical teams.

Practical tools bridge the gap between principles and action: The framework provides concrete resources like an Ethics Checklist and Portfolio of Measures that help translate high-level ethical requirements into specific technical and organizational actions, addressing the common challenge of knowing what ethics actually means in practice.

Training effectiveness improves through iterative refinement: The authors conducted pilot training sessions and documented lessons learned, suggesting that ethics competency frameworks benefit from real-world testing and adaptation rather than one-size-fits-all approaches.