“Will AI steal my glory?”: Power relations perceived by college instructors when grappling with Generative AI

Synthesis: Zuo, Xu & Dunning (2026) present an instrumental case study of 33 college instructors (10 interviewed in depth) exploring how their perceptions of Generative AI as a tool for student writing are shaped by four polycentric power centers: the global impact of AI, the university/department, colleagues, and students. Framed by Blommaert's notions of scale and polycentricity, the study shows how instructors navigate multiple, sometimes conflicting, centers of normativity — adjusting their teaching to retain authority and promote student accountability. It offers direct evidence on how instructor perceptions and practices evolve in writing instruction under GenAI, with implications for professional development.

Key Findings

1. Four polycentric power centers shape instructor perceptions. Instructors' views and practices are not uniform but shaped by their orientation to four centers of authority: the global impact of GenAI, the university/department, colleagues, and students — each explored in detail below.
2. The global trend is seen as inescapable. All instructors recognized GenAI as "the reality" — inevitable and "never going away" — motivating them to adjust teaching and "stay a step ahead" of students, though views differed on its effectiveness and ethics.
3. Instructors updated practices to retain authority and promote accountability. Across the four centers, instructors responded by emphasizing the writing process over the final product (reflections, more conferences, multiple revisions), holding open conversations about GenAI's limitations, and clarifying how GenAI may be used in each class.
4. The shift is from detection toward teaching acceptable use. Rather than relying on detection, instructors moved toward guiding responsible GenAI use and promoting student accountability — while flagging persistent equity concerns (the digital divide).

Theoretical framing: scale and polycentricity

The study applies Blommaert's notions of scale and polycentricity, rarely used in GenAI research, to analyze power relations in instructors' AI pedagogical approaches. Scale imagines social phenomena as "hierarchically ranked and stratified," across horizontal (spaces) and vertical (local-to-global) dimensions. Polycentricity holds that multiple "evaluating authorities" or centers coexist at different scales, and individuals adapt their practices to meet these centers' expectations. This explains why instructors' AI perceptions are non-uniform and shaped by conflicting normative centers.

Methodology

An instrumental case study (exploratory approach) at "Palm Tree University," a four-year U.S. liberal-arts institution that had formed a University GenAI Task Force and a Quality Enhancement Plan on information literacy. Data came from a questionnaire (33 of 265 faculty) and in-depth interviews with 10 focal participants across disciplines (English, Psychology, Education, Management, Biology, Medicine) and teaching experiences (3–27 years). Analysis used inductive coding (three rounds), thematic analysis across cases, and categorization/theorization.

The four power centers in detail

• Global trend of GenAI: The highest-level center; instructors felt they could not avoid the trend. Most saw affordances (brainstorming, drafting, support for language learners and students with learning disabilities, an alternative to expensive prep schools, grading time-savings), while some were concerned about limitations and ethical issues. The title's "Will AI steal my glory?" captures an instructor's distress at discovering ChatGPT could generate work meeting all her assignment criteria.
• University and department: Varying policies — some departments required syllabus AI statements while allowing instructor discretion; others had none. Instructors held mixed attitudes toward unified policy (wanting consistency vs. fearing restriction), and adjuncts felt limited by their discourse power.
• Colleagues: Perceptions spanned "absolutely the whole spectrum" — from senior instructors holding back to active advocates who saw ongoing learning as a departmental responsibility.
• Students: The direct stakeholders; instructors felt pressured to learn because "students will use it." Concerns included evaluating real proficiency, over-reliance ("laziness"), the Digital Divide (unequal access to better tools), and ambiguity about whether writing or digital literacy was being assessed.

What this means for practice

• Instructors. Emphasize the writing process over the final product — assign reflections on thinking, writing, and learning, require multiple revisions, and add instructor-student conferences — so that promoting student accountability does not depend on detection.
• Instructors. State the GenAI rules for each class explicitly in the syllabus and hold open conversations about the tool's limitations, moving from detection toward teaching acceptable use; departmental policy at the study site ranged from detailed AI statements to none at all.
• Faculty developers. Center professional development on process-oriented Pedagogies and Teaching Strategies and on open, cross-department dialogue about GenAI's affordances and risks, reaching senior colleagues who hold back and adjunct faculty whose limited discourse power constrains how they respond.
• Faculty developers. Build both faculty and student AI Literacy and address the Digital Divide and over-reliance directly, since instructors reported students with greater economic capital using better tools and using GenAI as a shortcut around learning.

Limitations

• Single-site instrumental case study at one four-year U.S. liberal-arts university ("Palm Tree University") that had already formed a GenAI task force and a Quality Enhancement Plan, so the four power centers may look different where no such institutional support exists.
• Questionnaire response was low (33 of 265 faculty), and the in-depth analysis rests on only 10 interviews, so the themes describe a small, self-selected group rather than the faculty as a whole.
• All evidence is instructors' self-reported perceptions gathered at a single time point in Fall 2024; the study documents what instructors say they do, not observed classroom practice or student learning.
• The 10 interviewees came from a fixed set of disciplines (English, Psychology, Education, Management, Biology, Medicine) with 3-27 years of teaching experience, leaving other fields and career stages unrepresented.

Citation

Zuo, Y., Xu, X., & Dunning, L. A. (2026). "Will AI steal my glory?": Power relations perceived by college instructors when grappling with Generative AI. Journal of Applied Learning & Teaching, 9(1).