Core Problem & Framework
The Problem: Generative AI can make schoolwork appear successful while bypassing interpretive labor, authorial ownership, and durable learning processes. Output looks like learning.
The Framework: Pedagogical friction distinguishes productive friction (which supports interpretation, authorship, and judgment) from exclusionary friction (which blocks access). The study examines three learner-facing dimensions: Noetic, Rhetorical, and Existential. Infrastructural friction names the active system-level constraints and supports that can sustain or weaken the other three; the institutional environment is not a neutral backdrop.
Research Questions & Methods
| Research Question | Participants & Instruments | Constructs & Analysis |
|---|---|---|
| RQ1: How do classroom-facing educators, building-level administrators, district or system-level leaders, and adult university students make sense of the ways generative AI changes the effort, dialogue, authorship, judgment, and participation involved in teaching and learning? | Interviews across all four participant groups. Card sorts only for classroom-facing educators and adult university students. Institutional documents and open-ended survey responses extend the qualitative evidence. | Noetic, rhetorical, and existential friction. Interpretive codebook thematic analysis, card-sort reasoning, descriptive survey context, and attention to discrepant accounts. |
| RQ2: How do educators and school-system leaders make sense of and respond to the policy, professional-learning, assessment, access, and governance conditions surrounding generative AI? | Educators and school-system leaders. Interviews, classroom-facing educator card-sort evidence where relevant, institutional documents, survey responses, and national structural context. | Infrastructural friction. Interpretive codebook thematic analysis, cross-role comparison, document analysis, and descriptive or conditional survey comparisons. |
| RQ3: What language, assessment expectations, instructional practices, professional supports, and leadership approaches do educators and school-system leaders identify as important for preserving forms of friction that support learning while reducing unnecessary barriers? | Educator and school-system leader evidence with supporting survey and national context. | Cross-role synthesis, joint-display integration, and meta-inference development through convergence, divergence, expansion, or silence. |
Primary Claims
1. Polished output cannot be assumed to demonstrate learning on its own.
Generative AI can reproduce markers associated with cognitive processing, so educators need additional ways to examine reasoning, authorship, and judgment.
2. Pedagogical friction is a design and governance concept.
Not all difficulty is productive; schools must govern which forms of friction to preserve and which to reduce.
3. Tertiary algorithmicity alters learning conditions.
Algorithmic generation shifts meaning-making from human interpretation to machine-shaped defaults.
Methodological Boundaries
- Scope: The study examines sensemaking among educators, school-system leaders, and adult university students. It does not measure direct student learning outcomes.
- Evidence hierarchy: Interviews and other open-ended evidence form the qualitative core. Closed-ended survey responses and national datasets provide supporting quantitative and structural context.
- Sampling: Approximately 6-8 classroom-facing educators, 3-4 building-level administrators, 3-4 district or system-level leaders, approximately 4 adult university students, and approximately 200 K-12 survey respondents. If the shared-core survey sample is below 80, analysis remains descriptive.
- Nonparticipant comparison: AI-generated texts and agentic artifacts remain analytically distinct, with at least one bounded, prespecified agentic-artifact task retained if other comparison elements are scaled.
- Generalizability: As an interpretive qualitative study, the findings may support reasoned transfer to similar contexts, but they do not provide universal statistical generalizability.
- Integrity vs. Learning: The study deliberately pivots from academic misconduct policies to focus on the preservation of learning processes.
Ethics, Governance & Research Status
- Current gate: Revise the interview-question wording, obtain committee review, rehearse the revised questions with three educators for conversational flow, and update the IRB materials. IRB approval and applicable site permissions precede formal recruitment or main-study data collection.
- Data security: Consent records, recordings, transcripts, survey responses, and identifiable research records belong only in encrypted, access-controlled, IRB-approved storage.
- Public-site boundary: GitHub Pages supports explanation and rehearsal only; it does not receive participant responses or participant data.
- AI boundary: No participant data will be supplied to generative or agentic AI systems. Comparison artifacts remain separately labeled nonparticipant records and are analyzed only after human themes are finalized.
- Researcher role: Primary participant recruitment occurs outside the researcher's district and through relationships in which the researcher holds no supervisory authority.