Illuminating complex student realities of artificial intelligence through an entangled pedagogy framework

Synthesis: Fawns, Bearman, Corbin, Henderson, Walton, Liang, McLean, Oberg & Matthews (2026) offer a theoretically informed qualitative analysis of how three university students (John, Vivian, Howard) negotiate generative AI within the "messy" realities of everyday study and life. Using Fawns' (2022) entangled pedagogy framework — grounded in Barad's sociomaterial/relational ontology — the study shows that GenAI engagement is inseparable from students' overlapping identities (learner, future professional, employee, parent), beliefs about learning, emotional responses, pragmatic constraints, trust relations, and moral judgments about responsibility, authenticity, and fairness. Student positions on GenAI are multiple, ambiguous, provisional, contextual, and sometimes internally conflicting. Contrasting with a sector fixated on staff development, Assessment redesign, and student compliance (e.g., AI detection), the authors argue policy and guidance should be co-designed with students and attend to lived realities rather than treating GenAI as a problem for universities. Students' positionings, they contend, are often more thoughtful and ethically attuned than public discourse suggests.

Key Findings

• Entangled pedagogy as an analytic lens. The study applies Fawns' (2022) framework, which treats technologies as integral to, rather than separable from, educational activity. Drawing on sociomaterial scholarship (Barad, Fenwick, Orlikowski), the authors treat GenAI engagement as emerging from shifting configurations of social, material, institutional, and personal relations — resisting instrumental or solutionist framings that abstract AI from context.
• Students hold multiple, overlapping identities. John is simultaneously a postgraduate student, a father, and a mathematics educator; Vivian is a neurodivergent (ADHD) art history student; Howard is an arts/psychology major working at a hardware store. Each negotiates GenAI across the intersections of these roles rather than as a discrete "student" decision.
• Positions are provisional, ambiguous, and sometimes conflicting. John criticizes peers "getting away with using AI" yet also criticizes universities for prohibiting it; Vivian rejects GenAI as inauthentic yet admits "AI sits across a whole lot of platforms... you can't avoid it entirely"; Howard trusts GenAI "to be untrustworthy" and uses it in ways that "technically do" breach rules while believing he upholds his personal standard of academic integrity.
• Emotion, trust, and morality are entangled with pragmatics. GenAI use is negotiated amid guilt, fear, frustration, and pragmatic pressure (e.g., Howard submitting assignments "with 2 min to spare," his fear of detection and "harsh consequences"). Decisions blend personal integrity, institutional expectations, trust relations with institutions/peers/technologies, and concerns about future employability and the "human element" of learning.
• Sector approaches mis-frame the issue. Dominant responses — staff development, assessment redesign, compliance regimes, AI detection — rely on students complying with policy despite evidence their views diverge from educator expectations. Policy does not translate cleanly into practice, because GenAI is entangled with ethics, identities, relationships, aspirations, and external pressures.

What this means for practice

• Educators. Create collaborative dialogic spaces where students can voice tensions about GenAI and make them visible "by design" (a seamful approach), rather than resolving complexity through top-down rules or surveillance.
• Educators. Foster reflective, ethical, context-sensitive judgment instead of treating AI Literacy as purely technical competency, and move past binary framings of GenAI as simply harmful or beneficial.
• Administrators. Replace compliance regimes and AI detection with cultures of openness, honesty, and vulnerability, positioning students as co-interpreters and co-designers rather than passive recipients of policy or potential rule-breakers.
• Administrators. Coordinate GenAI expectations centrally rather than leaving each subject coordinator to decide alone, since responsibility delegated to individual coordinators produces the inconsistency students must then navigate.
• Policymakers. Co-design institutional guidance with students and acknowledge that equity, diversity, and inclusivity make a "level playing field" hard to guarantee when students must exercise their own contextual judgment.

Limitations

• The analysis foregrounds only three participants (John, Vivian, Howard), and the authors state explicitly that they do not aim to generalize across the dataset, compare groups, or comment on prevalence.
• All three analyzed participants came from the same university, even though the wider focus group dataset spanned 79 students in 20 focus groups across four Australian universities.
• The study is a secondary analysis of existing focus group data rather than purpose-designed individual interviews, and the authors treat accounts as co-constructed within focus group interactions, so they are contingent on those conversational conditions.
• Participant selection was purposefully driven by the lead author's identification of analytically rich excerpts (40 identified, eight transcripts reviewed), which the authors present as one partial articulation of an always incomplete inquiry.

Citation

Fawns, T., Bearman, M., Corbin, T., Henderson, M., Walton, J., Liang, Y., McLean, J., Oberg, G., & Matthews, K. E. (2026). Illuminating complex student realities of artificial intelligence through an entangled pedagogy framework. Higher Education.