Inclusion and Equity as a Paradigm Shift for Artificial Intelligence in Education

Artificial intelligence (AI) tools and methods allow researchers and educators to assess complex patterns and relations among diverse variables (e.g., learner backgrounds, behaviors, learning contexts, and outcomes) to generate inferences and predictions for supporting learners, and to characterize learners at differing scales (e.g., from districts to individuals). As a paradigm shift, artificial intelligence in education (AIEd) experts are increasingly attending to questions of diversity, equity, inclusion (DEI) within these efforts. For example, there is growing awareness of algorithmic bias, such that algorithms and automated systems can recreate or exacerbate discriminatory or prejudicial outcomes. Similarly, conclusions based on statistical means can be misleading or exclusionary for learners who do not conform to “average” or majority demographics. To address such concerns, AIEd scholars must consider alternative approaches for studying educational phenomena, analyzing data, and drawing meaningful educational conclusions. Models may need to be disaggregated to include more nuanced variables and effects related to demographic factors and social identities. Simultaneously, intersectional approaches are needed to represent learners’ multiple identities (and associated power, privilege, and history), and to interpret observed effects. This paradigm shift is poised to empower more personalized and effective outcomes for a greater diversity of learners.