Abstract
The discourse of ‘learning crisis’ in the field of education and international development is frequently invoked by international institutions to inform policy and intervention strategies (Sriprakash et al., 2020). In this context, educational technologies are often optimistically hailed as potential solutions and agents of transformation (Teräs et al., 2020; Williamson, 2021). Increasingly, these educational technologies are shaped by the growing research field of AI in education (AIED), as exemplified by numerous algorithmic applications in education ranging from intelligent tutoring systems for personalised learning to learning analytics for student dropout prediction (Bustamante & Garcia-Bedoya, 2021; Celik et al., 2022; Chaudhry & Kazim, 2022). Nonetheless, evidence has demonstrated that despite the promised benefits, these technologies often not only fail to solve––but in fact exacerbate––the ‘learning crisis’ and the broader education crisis within a racialized and gendered technocapitalist system (Madaio et al., 2022). As such, there is an increasing emphasis on foregrounding discussions on the ethics of AIED, understanding its benefits and harms to enable effective governance and responsible use (Holmes et al., 2022).
In this paper, I approach AIED ethics through a decolonial intersectional feminist lens to question the unhistoricized and disembodied claims, frameworks, and practices underpinning the technocracy of AIED. Rather than considering AI systems in education as isolated technological objects, I recognize them as material, embodied, and entangled with the worlds and agencies they produce and are produced in (Bassett, 2023; Benjamin, 2019; D’Ignazio & Klein, 2020; Wilcox, 2023). Through historicizing and socializing AIED ethics within global racial and gender dynamics, I excavate how injustices in AIED operate both within the ‘crisis’ discourse in education development and within the data and algorithms underlying AI systems.