Educational AI: A Critical Exploration of Layers of Production and Productivity

Abstract

Regarding possible implications for teaching and learning, the article explores the production and productive effects of educational AI from sociology of knowledge/ of technology perspectives from three sides: Firstly, the role of knowledge (re-)construction in the creation of educational AI is investigated. In this context, contrasting engineeringoriented approaches, educational AI systems are conceptualised as agentic entities infused with tacit and explicit knowledge about sociality and education, and as potentially reshaping both educational practices and scientific concepts. Looking at promotional and engineeringoriented AI discourses, the article secondly examines how education and AI are linked and how the knowledge pervasion of educational AI is addressed. Findings indicate that the discursive production of educational AI relates to the interwoven assumptions that education and specifically lifelong learning are obliged and able to remedy large-scale societal challenges and that educational AI can leverage this potential. They also indicate that an educational AI system’s knowledge is deemed a reflection of explicit (expert) knowledge that in the form of rationales can, in turn, be reflected to the systems’ users. Thirdly, regarding arising challenges for the sensitive area of education, educational AI’s role in knowledge gathering practices both in educational research and big educational data analysis is addressed.