Ensuring Fairness Stability for Disentangling Social Inequality in Access to Education: the FAiRDAS General Method

Recent advancements in Artificial Intelligence in Education (AIEd) have revolutionized educational practices using machine learning to extract insights from students' activities and behaviours. Performance prediction, a key domain within AIEd, aims to enhance student achievement levels and address sustainable development goals related to education, health, gender equality, and economic growth. However, the potential of AIEd to contribute to these goals is hindered by the lack of attention to fairness in prediction algorithms, leading to educational inequality. To address this gap, we introduce FAiRDAS a general framework that models long-term fairness as an abstract dynamic system. Our approach, illustrated through a case study in AIEd with real data, offers a customizable solution to promote long-term fairness while promoting the stability of mitigation actions over time.