Explore: Assessment, teaching experience, and curriculum →

Teaching as accountable inquiry

I teach students to turn an experience that matters into a question they can investigate and an argument they can defend. A policy may appear untrustworthy or a representation troubling before a student has the vocabulary to explain why. I help students develop that vocabulary through close reading, conceptual explanation, evidence, and revision, while taking seriously the cultural knowledge that gave the question significance.

In communication, that responsibility is concrete. A reading of a media text, a coding decision, or a judgment about whether an institution deserves trust has consequences for the people it describes. My research asks when institutions remain answerable to the people who depend on them, and I hold my courses to the same test: students should be able to rely on them. Every staged task, from a research question to a proposal, an ethics application, a coded dataset, and a report, comes with a guide, a template, and a worked example, so that the structure of the course is visible before students need it.

Inclusive rigor

I hold every student to the same standard of defensible interpretation and vary the routes to it: worked examples, staged drafts, feedback on work in progress, and, in Philippine classrooms, explanation in both Filipino and English. I distinguish the productive difficulty of interpretation from unnecessary uncertainty about what an assignment requires. A standard that only some students can reach by the route offered is a weak test of what they know.

Knowledge and judgment before delegation

Generative AI makes foundational understanding more important, not less. Students need to recall concepts, explain arguments, read evidence, and recognize errors before they can reliably evaluate an automated answer. A fluent product cannot by itself show that its author has developed those capacities.

Students therefore meet primary readings and their own data first, annotating claims and evidence and attempting an interpretation before consulting any tool. Printed or otherwise markable readings support sustained attention, and accessible digital equivalents remain available. In face-to-face teaching I favor a device-down default, with set periods for purposeful use and exceptions for accessibility needs.

Generative AI as a conditional scaffold

I permit disclosed, assistive use of generative AI unless an activity requires unaided work, and always within institutional policy. Permission follows the learning objective. Some tasks build recall, reading, and initial reasoning without AI; others allow help with organization or feedback after a first attempt; in designated activities, students examine AI itself: what an output omits, how it frames a problem, and whether its evidence can be verified.

Where AI enters, it enters in sequence: the student's own explanation, then an AI critique, then a revision the student can justify. Students remain responsible for their sources, interpretations, and the work submitted in their name, and they disclose material assistance. Independence concerns responsibility and judgment; it does not require learning without support.¹ My research on how supported thinking is judged in academic life informs how I assess students' work.²

My own standard. The same rule applies to me. Every syllabus carries this statement:

I sometimes use generative AI tools to help me organize or phrase written feedback. I read every student's work myself, and every judgment in my feedback is mine. Before any feedback is returned, I check that it addresses the right student and the right text, that each comment refers to something actually in the work, and that I can explain and defend every point if asked. I do not use AI to grade. If you think a comment does not fit your work, tell me and I will review it with you.

Mentoring and supervision

Some students go further and present or publish their work. I treat this as an extension of a course, not a reward within it. Invitations to continue a project toward publication come only after final grades, are voluntary, are open to every team through a published route, and never affect the feedback a team receives. Co-authorship follows substantive contribution.

In doctoral supervision I set expectations in writing at the start and work through feedback memos, revision plans, and agreed timelines. I have supervised the doctoral research of Rae Francis Quilantang at the Polytechnic University of the Philippines, and our collaboration has produced co-authored articles, including two on teaching.³ Differences of language, neurotype, discipline, and regional formation can become resources for inquiry when expectations and routes to participation are explicit.

Teaching and research together

Teaching has educational purposes of its own: students need not adopt my research agenda, or produce material for it, to benefit from a course. My teaching and research have nonetheless developed together. Published work from my classrooms includes a two-session unit on multimodal critical discourse analysis built around The Boys, written for other instructors to adapt,⁴ and a set of standards for using generative AI in qualitative research, with exercises comparing students' coding with an AI's.⁵

The outcome I aim for is a student who can read carefully, ask consequential questions, evaluate evidence, use support deliberately, and remain responsible for a judgment.

Explore course materials → · Discuss teaching or mentoring

Notes

1. Karl Patrick R. Mendoza, “Cognitive Scaffolds, Not Caves: Neurodivergent Scholarship and Generative AI in the Accelerated Academy,” Disability & Society 41, no. 8 (2026): 2002–21, https://doi.org/10.1080/09687599.2026.2667539.
2. Karl Patrick R. Mendoza, “When Thinking Is Judged: Academic Ethics and the Moralization of Cognition,” Journal of Academic Ethics 24 (2026): article 81, https://doi.org/10.1007/s10805-026-09762-z.
3. Rae Francis Quilantang and Karl Patrick R. Mendoza, “Teaching in Entanglement: Co-thinking, Cognitive Justice, and Theory-Making in the Age of Generative AI,” Communication Teacher 40, no. 2 (2026): 250–52, https://doi.org/10.1080/17404622.2025.2609611.
4. Rae Francis Quilantang and Karl Patrick R. Mendoza, “Critiquing Power and Ethics: Teaching Multimodal Critical Discourse Analysis and Ethical Literary Criticism through Amazon Prime’s The Boys,” Communication Teacher 39, no. 4 (2025): 311–16, https://doi.org/10.1080/17404622.2025.2467804.
5. Karl Patrick R. Mendoza and Rae Francis Quilantang, “In Relation, Not Replacement: A Reflexive Case for Generative AI in Qualitative Research,” Qualitative Research Journal (2026), https://doi.org/10.1108/QRJ-11-2025-0428.