This study investigates how assessment practices are being re-imagined to align with Artificial Intelligence in Education (AIEd) pedagogical practices. As AI systems increasingly mediate learning through adaptive tutoring, co-creation, and metacognitive feedback, assessment frameworks must transform to become valid and pedagogically aligned. Combining Fairclough’s (1992) Critical Discourse Analysis (CDA) and the Theory of Practice Architectures (TPA) (Kemmis et al., 2014), this qualitative empirical study investigates how two Heads of Department based in Kazakhstan and Thailand construct meanings around AI-enabled assessment through their professional discourse. Data were collected via asynchronous email interviews and documentary artefacts evidencing institutional AI initiatives. Findings reveal that teachers discursively position AI as an assistant rather than an authority, anchoring their practices within dominant discourses of integrity, efficiency, and policy vagueness. CDA-TPA mapping demonstrates that teachers’ sayings highlight ethical vigilance, their doings show cautious experimentation constrained by infrastructural and linguistic inequities, and their relatings suggest hierarchical yet collaborative professional networks. However, students are often framed as suspect users of AI rather than reflective co-agents, thereby highlighting asymmetries in assessment relations. The study identifies a shift from regulative to reflective practice architectures, where AI begins to mediate trust, authority, and assessment design. It concludes that genuine alignment between AIEd pedagogy and assessment requires more cultural-discursive opportunities for reflective dialogue, material support for innovation, and policy frameworks that enable collaborative design. This paper contributes an original CDA-TPA synthesis for analysing power, ethics, and practice in AI-mediated educational assessment.

Authors:
Denise Germaine Taylor, Lancaster University, United Kingdom

About the Presenter(s)
Denise Germaine Taylor is a Doctoral Researcher in Technology Enhanced Learning at Lancaster University, focusing on Artificial Intelligence-aligned assessment in educational settings. Her general interests include travelling, reading about geopolitics and medical advances, attending theatre productions, and playing and listening to classical music. She is currently volunteering with the NGO AWC Staff Development in South Africa as a Technology Consultant, while researching AI in education with a sharpened focus on aligning AIEd practices with assessment.

Connect on Linkedin
https://www.linkedin.com/in/denian3368/

See this presentation on the full schedule – Saturday Schedule

Bookmark