Reframing Teachers as Co-Designers
of AI-Mediated Writing Pedagogy: Structuring Feedback Through P Versus NP
Conference website: https://baal2026.com/
Jerry Talandis Jr. – University of Toyama, Japan
Theron Muller – Waseda University, Japan
2 September, 2025, 13:00 – 13:30
Our Presentation
Abstract
Educational technologies for writing instruction are traditionally designed top-down, with pedagogical assumptions embedded by developers and delivered to teachers as fixed functionalities. However, in the context of generative AI (GenAI), this risks further marginalizing instructors amid broader transformations in writing pedagogy. Instead, effective and ethical GenAI integration in language education requires teachers to be co-designers. To support this, we introduce a pedagogical design framework grounded in the computational distinction between P (polynomial-time) and NP (nondeterministic polynomial-time) problem classes (Cook, 1971), demonstrating its implementation in a GenAI-enhanced L2 Writing curriculum. We differentiate between writing tasks that admit clear solutions (P) and those that involve open-ended reasoning (NP). While both are essential to writing development, they warrant different pedagogical responses: P-like tasks are suited to corrective, algorithmic feedback, whereas NP-like tasks require formative guidance to encourage reflection. We describe how we operationalized this framework so that instructors can configure task type, feedback modalities, and GenAI prompt constraints to determine when direct corrections, reflective questions, or student justification is required. This positions teachers as designers of GenAI behavior, enabling control over automation and human judgment, thereby offering an explicit ethical model for human–AI collaboration in language education. By restricting GenAI to supportive, dialogic roles in NP-like contexts and reserving evaluative authority for instructors, this system mitigates automation bias, protects student agency, and preserves writing as a cognitive activity. GenAI is thus framed as a pedagogically constrained collaborator embedded within teacher-determined design parameters rather than a surrogate teacher or authoritative evaluator. This presentation offers (1) a theoretically grounded design framework for GenAI-mediated feedback, (2) an operational model of teachers as co-designers within educational authoring tools, and (3) a practical demonstration of how ethical principles were instantiated. Together, these advance debates on teacher agency, responsible GenAI integration, and the transformation of pedagogically driven educational technologies.
References
Cook, S. A. (1971). The complexity of theorem-proving procedures. Proceedings of the Third Annual ACM Symposium on Theory of Computing, 151–158. https://doi.org/10.1145/800157.805047
Leave a Reply