How Colonial Continuities Underlie Algorithmic Injustices in Education

Several aspects of education, including student learning, testing, school assignments, and teacher evaluations are being impacted by a rapid increase in the use of data-driven and automated algorithmic systems. These systems have now come under scrutiny for their discriminatory behaviors that also closely resemble historical injustices. Yet, the research surrounding algorithmic fairness in educational systems conceptualizes this issue predominantly as a technical problem. We claim that there are severe limitations to this approach, especially if we want to shift our focus from the narrow and often problematic definitions of fairness to achieving equity and justice for the marginalized people. In this paper, we argue that the historical injustices perpetuated by these systems are the result of the colonial ways of knowing (and being) that continue to shape them. Our work asserts that in order to progress in the direction of building equitable systems in education, we need to critically examine the assumptions that underlie their design, development, and deployment. Using three illustrative scenarios, we socio-historically contextualize some of the fundamental principles of data and algorithms to reveal that they are still rooted in coloniality. We conclude by inviting researchers to intentionally question their practices and reflect on their position at the forefront of knowledge production on technological advancements in education.