Algorithmic Bias in Education

TLDR

Algorithmic bias in education has been studied mainly through fairness definitions, but this review focuses on concretely identifying which student groups are affected and at which stages of algorithm development and deployment bias occurs. The paper reviews causes and documented manifestations of algorithmic bias in education and proposes a framework to transition from unknown to known bias and from fairness to equity, along with four mitigation strategies. The authors synthesize theoretical perspectives, connect them to the machine learning pipeline, evaluate bias metrics, and examine evidence across race/ethnicity, gender, nationality, socioeconomic status, disability, and military‑connected status, while outlining mitigation efforts. They present a framework that moves from unknown bias to known bias and from fairness to equity, offering a structured approach to addressing algorithmic bias in educational technology.

Abstract

In this paper, we review algorithmic bias in education, discussing the causes of that bias and reviewing the empirical literature on the specific ways that algorithmic bias is known to have manifested in education. While other recent work has reviewed mathematical definitions of fairness and expanded algorithmic approaches to reducing bias, our review focuses instead on solidifying the current understanding of the concrete impacts of algorithmic bias in education—which groups are known to be impacted and which stages and agents in the development and deployment of educational algorithms are implicated. We discuss theoretical and formal perspectives on algorithmic bias, connect those perspectives to the machine learning pipeline, and review metrics for assessing bias. Next, we review the evidence around algorithmic bias in education, beginning with the most heavily-studied categories of race/ethnicity, gender, and nationality, and moving to the available evidence of bias for less-studied categories, such as socioeconomic status, disability, and military-connected status. Acknowledging the gaps in what has been studied, we propose a framework for moving from unknown bias to known bias and from fairness to equity. We discuss obstacles to addressing these challenges and propose four areas of effort for mitigating and resolving the problems of algorithmic bias in AIED systems and other educational technology.