Supporting assessment redesign in the age of AI - a critical practice-based framework for equitable change

Synthesis: This paper argues that generative AI necessitates reimagining higher education Assessment beyond integrity concerns toward critical, equity-focused redesign. Prue Laidlaw (Charles Sturt University) develops a theoretical framework integrating the Theory of Practice Architectures with critical pedagogy to guide AI-integrated assessment reform, informed by participatory action research in an Australian university. Assessment redesign is conceptualized as a multidimensional practice requiring simultaneous attention to three dimensions: cultural-discursive elements (the language and assumptions that shape practice), material-economic arrangements (resources, tools, infrastructure), and social-political factors (governance and power). A four-phase implementation matrix — critical analysis, intervention design, equity-focused implementation, and evaluation — operationalizes the framework, positioning educators as critical agents shaping AI integration rather than passive recipients of technological change.

The central argument is that technical solutions alone are insufficient: sustainable redesign requires cultural-discursive, material-economic, and social-political change to occur simultaneously, and when these dimensions remain misaligned — particularly when resources flow through existing power structures without explicit equity attention — even well-intentioned initiatives risk amplifying existing inequities. The framework identifies the resource-power intersection as both the greatest leverage point and the greatest risk for reform. In practice, successful redesign required participatory governance, critical AI literacy development, and explicit attention to resource-power intersections to maintain authenticity and pedagogical integrity. The author explicitly values incremental change, critical literacy development, and productive tension as legitimate outcomes, challenging narratives that privilege rapid, frictionless transformation.

The work connects to equity in AI education by making power and resource distribution — rather than detection or tooling — the center of assessment reform, and to assessment validity by reframing what a valid AI-era assessment is in terms of educational purpose rather than format. It resonates with authentic assessment and process-oriented assessment as design responses, while its practice-architectures lens adds a governance dimension those literatures treat less centrally. The framework also speaks to AIED institutional strategy, where the gap between prohibition and enthusiastic integration is exactly the incoherence it sets out to address.

What this means for practice

• Higher-education institutions and program leaders. Do not treat AI-era assessment redesign as a tooling or detection problem. Run all three dimensions at once: examine the language and assumptions in use, map where tools and resources actually flow, and inspect who holds decision-making power before changing any format.
• Use the four-phase matrix as a program checklist: (1) critically analyze current practice across all three dimensions, (2) co-design interventions with shared language, equitably accessible tools, and inclusive governance, (3) implement with differentiated support and active monitoring of power dynamics, and (4) evaluate discourse shifts, tool effectiveness across populations, and whether governance changes disrupted or reinforced hierarchies.
• Build critical AI literacy and participatory governance before deployment. The practice insights show that redesign stalls or backfires when institutions move resources through existing power structures without centering traditionally marginalized voices.
• Value incremental change and productive tension as legitimate outcomes. The framework explicitly rejects "frictionless implementation" narratives, which matters when leadership pressure demands rapid, uniform rollouts.

Limitations

• The framework emerged from a single Australian institutional context; the participatory action research means the empirical grounding is illustrative rather than validating, and the author calls for rigorous testing across diverse settings.
• Student voices are not yet centrally represented, and the author flags the need for future work to engage non-Western epistemologies for global applicability.
• The detailed implementation matrix is presented in the paper's Appendix A, so the strategies in this article's summary are the overview level rather than the full specification.
• As a theoretical and practice-informed framework rather than an outcomes study, the paper does not report measured effects on student learning, equity metrics, or assessment quality.

Citation

Laidlaw, Prue (2026). Supporting assessment redesign in the age of AI: a critical practice-based framework for equitable change. Teaching in Higher Education, 31:7, 1412–1435.