Reshaping higher education designs and futures: Postdigital co-design with generative artificial intelligence

Synthesis: Nineteen authors and three open reviewers ask what happens to co-design — the inclusive, iterative, relational practice of designing curricula, environments and institutions together — when GenAI is treated not as a tool but as a collaborator in the design process. The article is structured in three parts: a literature review of co-design, GenAI and postdigital theory; eleven situated contributions from educators, researchers and designers in ten countries; and a discussion of implications for higher education's design futures. Its central move is to distribute agency: where most AI-in-education work asks whether the technology helps or harms, this asks what design practice becomes when knowledge production, narrative construction and iterative Reflection are co-produced with a non-human participant. Two claims run through it — that GenAI simultaneously pluralises perspectives and defers to normative ones, so language fluency must not be read as breadth of knowing, and that the entanglement intensifies rather than removes the need for human interpretation and care.

This page summarizes a collective forum article rather than a study. It reports no experiment and collects no data: the evidence it offers is nineteen authors' practice-based accounts, plus three open reviews published alongside them.

How the Article Is Built

The first part reviews co-design's roots in participatory design, democratic engagement and distributed decision-making, and argues that co-design in higher education is necessarily postdigital because it already runs through a network of technology-mediated communication and design media. Human–AI collaboration then enters that network: GenAI is implicated in authorship, participation and the Ethics of who gets to shape a design.

The second part is the article's distinguishing feature — eleven short contributions, each by a named author or pair, each a situated account of practice rather than an evaluation. They include provoking critical engagement with an unreliable AI collaborator (Webster and Chubb), reimagining Chinese literature through postdigital co-design, GenAI as co-tutor (Hrastinski and Engwall), including GenAI in co-designing a postgraduate program (Mantai and Casey), cognitive entanglements in designing assessment frameworks, a "learning safari" rethinking faculty development, GenAI as co-actor in empathic co-design (Smeenk and van Dijk), GenAI as objects-to-think-with in design education, GenAI as a proxy participant in student co-design projects (Paulovich and Muñoz), and future-focused learning in digital graphics courses.

The third part, plus three open reviews (one on universities as civic partners, one on agency as endelecheia, one on co-design and the purpose of education), supplies the contestation that a single-authored position paper would not.

The Unreliable Collaborator: A Three-Phase Sequence

The most directly reusable contribution is Webster and Chubb's argument for treating GenAI's unreliability as design material rather than a defect to work around. Because AI-generated content is aggregated from human experience and its selection process is opaque, participants must critically assess how knowledge is selected, framed and contextualized — and because GenAI can homogenize imagination through normative training patterns even while appearing to offer alternative worldviews, its errors and inconsistencies become objects for critique.

They propose a three-phase approach, presented as successive but meant to be used flexibly:

• Provocateur. Groups frame an initial wicked problem, then treat GenAI outputs as provocations to be interrogated: whose voices are included or omitted, which ideologies are reinforced, what the group would reframe to reflect its shared priorities.
• Pluraliser. Participants co-design artifacts capturing different framings of the same issue by culture, class or gender, working iteratively with GenAI to generate counter-narratives and notice which perspectives the system defaults to.
• Facilitator. Groups reflect collectively on both the content and their own process, using the system to hold the deliberation rather than to answer it.

This sequence is the article's practical contribution to instructional design: the AI's role is deliberately weak in every phase, and the artifact that gets designed is a shared understanding rather than a product.

What the Collective Argument Amounts To

The conclusion extends co-design's ontology rather than reporting an outcome. Educational design, it argues, is relational: knowledge, values and practices emerge through ongoing human–technology interaction, so co-designing futures means engaging how knowledge is produced and valued in localized yet entangled contexts. GenAI's involvement does not erase the need for human interpretation, judgment and care — it intensifies them, which relocates the design question from what an entity possesses to what it does, and from individual authorship to co-agency.

The article is also candid about the tensions this creates. Several contributions keep human control firmly in place, and the collective reading is that this is not a failure of imagination but the current state of practice. Institutional change runs through the same ambivalence: GenAI can pluralize a collaboration and simultaneously defer to dominant ways of knowing, and the authors treat "recognizing these tensions" as the discipline that keeps postdigital co-design from collapsing into technical optimization.

What this means for practice

• Instructional designers. Use GenAI's limitations as the content. The provocateur–pluraliser–facilitator sequence works because the AI is asked for material to critique, not for a finished design, and the group's negotiation is what produces the artifact.
• Faculty developers. Treat the eleven contributions as a case library rather than a model. Each is a situated account from a different institution and country, which is more useful for adapting a co-design studio than a single transferable method would be.
• Administrators. Expect co-design with GenAI to distribute authorship, and govern accordingly. The article's caution is that language fluency is not breadth of knowing, which matters when AI-mediated design documents start carrying institutional authority.
• Researchers. Note the evidentiary status you are working with: a forum of practice accounts plus open review, not a study. It generates hypotheses about distributed agency and about normative homogenization; it does not test either.

Limitations

• There is no data, no intervention and no comparison. The eleven contributions are self-reported practice accounts, varied in length and in the depth of evidence they bring.
• The article is a forum with nineteen authors and three open reviews, so its argument is deliberately plural and uneven; a reader looking for a single position or a tested framework will not find one.
• The three-phase sequence is a proposal. No implementation, evaluation, or student outcome is reported for it.
• Contributors cluster in design, architecture, education and information-systems departments across largely Anglophone and Northern European institutions, so the practice base is narrower than the geographic spread suggests.
• The postdigital framing does real analytic work but is contested; the article argues from it rather than establishing it, and alternative accounts of the same phenomena are not considered.

Citation

Zeivots, S., Casey, A., Winchester, T., Webster, J., Wang, X., Tan, L., et al. (2025). Reshaping higher education designs and futures: Postdigital co-design with generative artificial intelligence. Postdigital Science and Education, 7, 1334–1374.