Metacognition as Disciplinary Infrastructure in AI-Mediated Learning
Source: International Journal of Emerging and Disruptive Innovation in Education: VISIONARIUM
Author: Tina R. Austin; Jason Gulya; Nick Potkalitsky
Original source: https://digitalcommons.lindenwood.edu/ijedie/vol4/iss1/6/
DOI: 10.62608/2831-3550.1055
Published: 2026-06-22
Source type: journal article
Summary
Austin, Gulya, and Potkalitsky argue that metacognition in AI-mediated learning cannot stay at the level of generic “thinking about thinking.” Because AI tools can take over parts of the cognitive work, students need discipline-specific ways to decide what to delegate, what to critique, and what to protect. The article frames metacognition as disciplinary infrastructure: students can only monitor their thinking well when they understand what their subject area requires of evidence, validity, interpretation, authorship, method, and care.
The article’s practical contribution is a checkpoint model for AI use. Students pause before, during, and after AI interaction to decide whether AI belongs in the task, how to evaluate AI output against disciplinary standards, what needs to be revised or remade, and what cognitive work should remain their own. The authors connect this to Austin’s UnBlooms framework and Gulya’s pre-creation loop, arguing that productive friction can deepen learning when it is timed to make disciplinary judgment visible rather than merely interrupting the flow of work.
Pull quotes
Metacognition needs disciplinary grounding
“Metacognition cannot remain a generic request for students to reflect. It must be anchored within a disciplinary infrastructure.”
AI use depends on selection, not just offloading
“The central pedagogical question, then, is not just whether students are offloading cognition, but how they decide what gets offloaded.”
Dual-system vigilance
“Traditional metacognitive frameworks ask learners to monitor one cognitive system (their own). AI-mediated learning introduces a second: one that produces fluent, confident output without genuine understanding.”
Resistance as part of AI literacy
“Educating people about the use of AI, then, involves teaching them to decide when and for what purpose it ought not to be used.”
Big ideas
- AI literacy has to be taught inside real subjects
- Learning still needs some struggle, even when AI can make things easier
- AI simulations need clear boundaries for learning
Claims
- AI literacy only works when it is connected to subject-area knowledge
- Students need boundaries for when to use AI and when to step back
- Learning requires some productive struggle that AI can remove
- In an AI world, assessment should focus on watching students think
Key evidence and examples
- The authors argue that students need metacognitive checkpoints before, during, and after AI use, not only end-of-assignment reflection.
- Austin’s UnBlooms framework is presented as a recursive sequence of questioning, generating, critiquing, and refining, with metacognitive reflection anchoring each move.
- Gulya’s pre-creation loop asks students to map a project process in advance and name where AI might help or harm ownership of the work.
- The article connects AI-mediated learning to disciplinary literacy: history, science, and writing each require different standards for evidence, validity, interpretation, and revision.
- The authors use the language of resistance to describe the principled decision not to use AI when the cognitive struggle itself is the learning goal.
Education relevance
This is highly relevant to AI literacy, disciplinary literacy, assessment redesign, writing instruction, student metacognition, and classroom routines for deciding when AI should assist versus when students need to do the thinking themselves.
Durability note
This source is durable because it ties several recurring wiki concerns together: disciplinary AI literacy, visible student thinking, productive friction, cognitive offloading, and student ownership of AI-mediated work.