Explainable AI in Education: Fostering Oversight and Shared Responsibility

This report emphasizes the vital role of Explainable AI (XAI) in education, focusing on how it can support the development of specific skills and competences across a wide range of stakeholders, to promote responsible, transparent, and critical engagement with AI systems. For the general public and end-users, there is a clear need for education and training to build the skills necessary to interpret AI explanations, ask relevant questions, and assess system trustworthiness, ultimately enhancing individual agency and critical thinking. For students, XAI should empower self-regulated learning by supporting metacognition, planning, and reflection.

Educators, in particular, require targeted AI and XAI competences, not only to understand and critically evaluate AI outputs but also to explain their use of AI tools in ethical and transparent ways. The report outlines a detailed framework of core educator competences, including the ability to navigate different dimensions of AI explanations, such as scope, depth, alternatives, and flow. Additionally, the skill of transparency is framed as an ethical imperative, enabling educators to justify their pedagogical choices involving AI to learners, parents, and colleagues.

For educational stakeholders more broadly, the emphasis is on cultivating a balanced set of knowledge, skills, and attitudes that enable informed decision-making and uphold human oversight in AI-supported environments. The text also distinguishes between different tiers of AI users, from novices who need basic understanding for trust and fairness, to advanced users requiring tools to evaluate and adjust models, and AI experts focused on model interpretability and improvement.

In conclusion, the text underscores that implementing XAI in education goes beyond technology; it requires a systemic effort to develop AI literacy and critical thinking as core educational goals, ensuring all users are equipped to engage meaningfully and responsibly with AI systems in ways that support trust, transparency, and ethical use.