Bibliographie sélective OHADA

The opacity of artificial intelligence (AI) technologies raises pressing ethical concerns in education, where algorithmic decisions impact learners’ rights, opportunities, and agency. These concerns may be addressed through transparency, fairness, and accountability (TFA) principles: transparency via disclosure of data, logic, and outputs, and the provision of understandable explanations; fairness through preventing unjust bias; and accountability through clear responsibility, oversight, and redress mechanisms. Explainable AI (XAI) clarifies decision-making processes behind AI outputs, supporting ethical governance in educational settings. This paper examines the legal and technical dimensions of XAI in education, focusing on the EU AI Act and the GDPR. It argues that explainability is a fundamental requirement for high-risk educational AI and identifies approaches to operationalise XAI requirements embedded in these frameworks. It further considers stakeholder-tailored explanation methods to bridge gaps in AI literacy and domain expertise. A TFA thematic analysis of three hypothetical use cases — automated grading systems, AI detection, and personalised learning applications — illustrates the interaction between legal obligations, technical methods, and educational practice. By mapping regulatory requirements onto technical strategies, the paper proposes a compliance framework for implementing explainability in educational AI, contributing to the emerging discourse on XAI by clarifying legal standards, technical approaches, and their implications for ethical governance.