Learning Management Systems (LMSs), predictive dashboards, and AI tutors have become so embedded in the digital education landscape that their presence often goes unquestioned. They are rolled out as inevitable, even desirable, responses to a changing educational world. As a result, educators and institutions alike frequently treat them as neutral infrastructure - tools that support teaching and learning without shaping it in return. But what if this sense of neutrality is precisely the problem?

This post begins with a simple provocation: what imaginary of education do these tools carry? What vision of teaching, learning, and institutional purpose is being embedded in their design, logic, and everyday use?

To explore this question, we draw on the work of philosopher Cornelius Castoriadis, who argued that all human societies are instituted through what he called the social-historical. For Castoriadis (1997), institutions are not fixed or eternal. They are created through collective acts of meaning-making. They embody a society’s shared imaginaries, its sense of what is real, what is possible, and what is legitimate. These imaginaries are not abstract ideas; they materialise in concrete structures, norms, and routines. Institutions naturalise these significations by embedding them into everyday practice. In doing so, they shape the horizons of what can be thought and done.

Education, from this perspective, is not merely a system for transmitting knowledge. It is an imaginary institution that carries, and reproduces, deeper social understandings of authority, autonomy, progress, and success. Digital education technologies do not escape this logic. They are not blank slates awaiting pedagogical input. They are already inscribed with specific visions of learning and power. These visions are often aligned with managerial logics, privileging measurement over meaning, automation over dialogue, and optimisation over exploration.

As Selwyn (2022) argues, educational technologies are shaped by cultural, economic, and political forces that embed assumptions about what learning is and how it should be governed. Technologies are not “just tools” but sites of power and ideology. They frame the problems we seek to solve and limit the kinds of solutions we can imagine. What appears as innovation may actually be a reproduction of dominant social norms under the guise of technical improvement.

This framing effect is rarely made explicit. Instead, the defaults of EdTech platforms operate as what Castoriadis would describe as instituted meaning. They present one particular version of education as if it were the only one. In doing so, they foreclose alternative futures before they can even be conceived.

If we are to reclaim digital education as a site of democratic possibility, we must begin by questioning its invisible frames. This means recognising that technologies are not neutral but saturated with social meanings. It means engaging with the institutional imaginaries they reproduce, and the imaginaries they exclude. Most importantly, it means seeing ourselves as agents within the social-historical, capable of instituting, questioning, and transforming the systems we inherit.

What is an institution? For Cornelius Castoriadis (1997), institutions are not fixed entities but socially created forms that organise collective life. They emerge from the radical imagination, the human capacity to posit new significations and structures, and become stabilised through the reproductive imagination, which sustains and naturalises what already exists. This distinction is crucial: institutions are not eternal, but contingent. They carry meanings that shape how we live, think, and act together.

Castoriadis describes institutions as expressions of the social imaginary, the shared meanings that a society constructs to make sense of itself. These meanings, what he calls imaginary significations, are embedded in institutional forms and practices, often becoming taken for granted. Institutions acquire a sense of “naturalness” not because they are inevitable, but because they are sustained through ritual, repetition, and social belief. As such, institutions appear not only in formal structures like schools or governments but in our tacit understandings of what counts as “normal” or “valid” behaviour within a given domain.

In education, institutions include not just buildings and bureaucracies but curricular structures, assessment conventions, learning management systems (LMSs), and the very idea of what constitutes “learning”. Consider how a traditional lecture hall encodes an imaginary of education as transmission from expert to passive recipients. Or how LMS course templates often replicate managerial logics of control, sequencing, and surveillance, privileging content delivery and analytics over co-creation and dialogue.

These are not neutral tools or spaces. They materialise and sustain particular imaginaries of what education is and should be. Even the concept of the “module” or “learning outcome” reflects institutionalised assumptions about the measurability and modularity of learning, echoing broader neoliberal imperatives of efficiency, accountability, and instrumentalisation (Selwyn, 2022).

Institutions play a dual role: they enable, and they constrain. They provide shared frameworks for meaning-making and collective action, enabling stability, social coordination, and identity formation. Educational institutions give shape to the idea of school, university, discipline, qualification. They offer continuity across generations, making possible what we might call “education” in the first place.

Yet they also impose boundaries. They delimit what can be thought, said, or done. They regulate behaviour, often marginalising alternative practices or forms of knowledge. For example, while LMS templates make course creation more efficient, they may also discourage pedagogical experimentation or learner-led structures. Similarly, traditional timetables and assessment calendars can limit opportunities for slow, reflective, or collaborative learning.

This ambivalence is not a flaw, but a fundamental feature of institutional life. Castoriadis does not advocate the abolition of institutions but calls on us to recognise their constructedness, and therefore their transformability. We must remain attuned to how institutions both open and foreclose possibilities, and to our role in either reproducing or reimagining them.

As educators, even modest gestures can be acts of radical imagination. Choosing to co-design assessment criteria with students, experimenting with ungraded feedback, or challenging default settings in a digital platform, these are not simply technical tweaks but interventions in the institutional imaginary. They signal that another education is possible, not by rejecting institution, but by renewing it.

Learning Management Systems (LMSs) such as Canvas, Moodle, and Blackboard have become the dominant infrastructure of digital education. Their ubiquity lends them an air of inevitability. Yet this apparent neutrality is itself a product of what Castoriadis (1987) terms the social imaginary, the shared, largely unspoken system of meanings that institutions carry and reproduce. These platforms are not blank slates; they encode a particular imaginary of what education is, or ought to be.

At first glance, an LMS might seem like a helpful administrative layer, a place to upload slides, host forums, or manage submissions. But the deeper architecture of these platforms reflects and reinforces specific pedagogical assumptions. Course content is broken down into weeks or modules, assessments are aligned to outcomes, and participation is tracked through analytics dashboards. The learning journey is presented as linear, modular, and measurable. These defaults are not merely conveniences, they are inscriptions of a technocratic vision of education: efficient, standardised, and accountable.

This vision is not imposed coercively, but becomes naturalised through what Castoriadis would describe as institutional repetition. When we log in to a platform and see the same structure, weekly folders, quizzes, slides, and rubrics, we begin to internalise the idea that this is what education looks like. And for educators pressed for time or support, the templates on offer seem like pragmatic solutions. Over time, the LMS ceases to be a tool and becomes the terrain itself.

The effects of this are subtle but profound. As Morris and Stommel (2018) argue, the LMS is not neutral. It reflects the pedagogical values of its designers. Features like auto-marked quizzes, learning pathways, or discussion boards are not just functional; they promote a transactional view of learning. The design of the platform encodes a default pedagogy, one that privileges clarity, consistency, and control over ambiguity, exploration, and dialogue.

These defaults operate as a kind of hidden curriculum, shaping not only what is taught but how teaching itself is understood. Teachers are positioned as content managers, students as users or consumers, and learning as a process of content acquisition and performance tracking. Selwyn (2022) notes that such systems reflect a broader managerial logic in education, one that seeks to render teaching visible, accountable, and auditable. This may support certain administrative imperatives, but it also narrows the space for more speculative or dialogic forms of pedagogy.

Crucially, this default setting marginalises approaches that do not fit neatly within the LMS paradigm. Pedagogies that prioritise inquiry, co-creation, or open-ended dialogue often require creative workarounds or external tools. Even speculative pedagogies, which invite learners to imagine and shape alternative futures, struggle to find space within systems designed for alignment, tracking, and metrics. In this way, the LMS acts not just as a repository or interface, but as an institution that closes down the radical imaginary, the capacity to imagine education otherwise.

Educators have not passively accepted these constraints. Some have resisted by bending the tools toward more dialogic or participatory ends: using discussion boards for collaborative storytelling, embedding external tools for creative expression, or treating the LMS as a scaffolding rather than a container. But these resistances often occur despite the platform, not because of it.

Understanding the LMS as a social-historical construction, one that carries and enforces particular imaginaries, is not to reject technology wholesale. Rather, it invites us to critically engage with the defaults we inherit and to ask what kinds of teaching and learning are being enabled or foreclosed. As this series explores, the work of reclaiming pedagogy begins with naming the structures that have come to seem inevitable, and imagining what might be otherwise.

Artificial intelligence (AI) is increasingly positioned as a solution to long-standing challenges in education, offering personalisation, efficiency, and real-time responsiveness. From predictive analytics and adaptive learning platforms to automated essay scoring and chatbot tutors, AI tools promise to tailor the learning experience to each student’s needs. Yet these promises often obscure the deeper pedagogical and political questions embedded in their design.

At their core, many AI systems operationalise a narrow vision of education, one that privileges quantifiable outcomes, measurable progress, and individual optimisation. While such tools may appear neutral or even benevolent, they carry embedded logics that shape how education is experienced and valued. Predictive analytics, for example, often use historical data to anticipate student “success,” but these models may reflect and reinforce institutional biases, penalising non-conforming trajectories. As Eubanks (2018) argues in her broader critique of algorithmic decision-making, these systems tend to “punish the poor” by automating exclusion under the guise of fairness.

Adaptive learning platforms similarly claim to personalise learning paths, yet their architecture often reduces knowledge to micro-tasks aligned with behavioural metrics. Tools like Knewton or Coursera’s recommendation engines exemplify this logic: learning becomes a series of checkboxes, and deviation from the norm is treated as inefficiency to be corrected. While these tools purport to individualise learning, they tend to standardise pedagogy, embedding what Knox et al. (2020) call the “automated teacher” into the infrastructure of instruction.

Grading tools such as Gradescope or Turnitin’s AI-assisted feedback similarly reshape assessment practices. These platforms offer efficiencies, but they also narrow the pedagogical relationship between teacher and student. Feedback becomes mechanised, and formative dialogue risks being replaced by algorithmic interpretation. This is not just a technical change, it is an epistemological one, redefining what counts as understanding or achievement.

Despite claims to innovation, these systems often reinforce managerial logics already dominant in higher education, control, surveillance, optimisation. As Selwyn (2022) notes, AI in education tends to align with institutional priorities: retention, progression, employability. Personalisation thus becomes a strategy not of emancipation but of compliance, where students are nudged toward predefined goals under the veneer of choice.

The imaginaries encoded in AI systems are rarely those of students or educators. Instead, they reflect the assumptions of developers, vendors, and policymakers, actors whose visions of learning are shaped by markets, metrics, and managerialism. When AI systems automate pedagogy, they also automate a specific vision of what education is for.

Castoriadis (1987) reminds us that heteronomy occurs when a society forgets its capacity to create its own institutions and instead treats inherited forms as given. AI risks accelerating this forgetting: presenting systems as inevitable, data-driven, and beyond contestation. But imagination, radical imagination, remains possible. Rather than reproducing inherited scripts, education can become a site for collectively imagining new ways of learning, ways that cannot be predicted or predetermined by data models.

Educators are not powerless within this landscape. They can intervene, by questioning system assumptions, by highlighting what is excluded, and by designing practices that foreground relational learning, critical dialogue, and open-ended inquiry. These gestures may be small, but they are acts of institutional imagination, reminding us that the future of education is not yet written.

Digital infrastructure does not simply support pedagogy, it shapes it, often by embedding assumptions about what learning should be. These assumptions, once coded into platform defaults, are no longer debated but enacted, making ideology function as design logic. In this way, educational infrastructure becomes a material expression of institutional priorities, not a neutral scaffold for delivering content.

Critical theorist Andrew Feenberg (2002) reminds us that technologies are never just tools; they are social artefacts shaped by the values and power dynamics of those who design them. Infrastructure encodes decisions, about what can be done, by whom, and how. Learning management systems (LMSs), analytics dashboards, and content delivery networks each make specific assumptions about how knowledge should be organised, accessed, and assessed. As such, infrastructure is not separate from pedagogy; it already teaches.

This entanglement is why digital infrastructure must be understood as a site of political struggle. Paulo Freire (1970) famously argued that all education is political, never neutral, and that every educational act either reinforces or challenges existing social orders. Infrastructure, too, participates in this process. A platform that limits student-to-student interaction, that privileges surveillance over dialogue, or that renders instructors as content managers rather than co-learners, is not merely “efficient”, it encodes a vision of education rooted in compliance and control.

Martin Weller (2020) highlights how the adoption of educational technologies is often accompanied by a narrative of inevitability, in which certain infrastructures are framed as both modern and necessary. But such narratives obscure the ideological work these systems perform. What appears to be progress may actually be entrenchment: a narrowing of pedagogical possibilities disguised as innovation. As Weller notes, edtech history is marked by repeated cycles of hype, adoption, and disappointment, not because the tools are inherently flawed, but because their pedagogical assumptions are rarely interrogated.

Understanding this shaping power requires a pedagogical lens, one that sees not only what infrastructure does, but what it presumes. The default features of widely adopted LMSs often reflect a managerial view of education: teaching is broken into modules; learning is tracked through rubrics and analytics; and student engagement is measured by clicks and timestamps. These defaults frame pedagogy as transactional, standardised, and optimisable. Even when intentions are more progressive, the infrastructure can nudge practice back toward these dominant norms.

To reclaim digital infrastructure as a site of democratic possibility, we must be willing to question these defaults, repurpose tools, or build alternatives. Open-source and federated platforms like Moodle and PeerTube provide instructive examples. Unlike proprietary systems, such tools can be adapted, reconfigured, or resisted by the communities who use them, supporting an ethos of autonomy and mutual responsibility. Their openness makes them more than platforms; they are possibilities, spaces in which pedagogy can be negotiated rather than dictated.

This does not mean every institution should abandon mainstream systems. But it does mean that educators, designers, and technologists must remain alert to the ideological work that infrastructure performs. Infrastructure is never just a backdrop. It is a curriculum. It teaches us what counts as legitimate practice, what kinds of participation are recognised, and what forms of imagination are rewarded or discouraged.

If infrastructure already teaches, what kinds of futures does it make possible, and which does it foreclose?

If institutions are made, they can also be remade. This deceptively simple claim sits at the heart of Castoriadis’ project, and at the heart of any emancipatory pedagogy. Education is not merely shaped by institutions; it is itself an institutional process, a continual interplay between form and imagination, reproduction and creation.

To imagine the institution otherwise means to recognise that the infrastructure of digital education, its platforms, templates, policies, workflows, and data regimes, is not a neutral container. It is a structuring force that sets the terms of what teaching and learning are understood to be before any educator logs in. Educational infrastructure, like any institution, reflects decisions about what is valued, who gets to decide, and how knowledge is organised and experienced.

Educators, therefore, must not only work within the infrastructure but also on it. This entails treating educational infrastructure not as passive delivery mechanism but as a site of critical engagement and possibility. As Morris and Stommel (2017) remind us, every instructional decision is a political one, and every platform embeds pedagogical choices that must be interrogated.

Reclaiming infrastructure might begin with small gestures. For example, an educator might subvert the linear logic of an LMS by structuring content around thematic inquiry rather than rigid weeks. Or they might disable completion tracking to signal that learning is not a checklist but a dialogue. Such acts may seem modest, but they resist the automation of pedagogy and reassert its creative, relational core.

To echo Biesta (2005), the point of education is not simply learning, but subjectification, socialisation, and qualification, held in tension and directed towards democratic life. Freire (1996) would call this an act of naming the world, a refusal to be scripted by systems, and a commitment to humanising pedagogy. The critical educator is not just a responder to infrastructure, but a participant in its ongoing creation.

Infrastructure may appear as silent plumbing, behind the walls, unseen and fixed, but in education, it scripts what is possible before teaching begins. If we do not engage in its critique and reconfiguration, we risk allowing educational futures to be dictated by market imperatives and algorithmic logic. Reimagining the institution is not utopian fantasy; it is a necessary pedagogical task.

The next post will explore the creative dimensions of critique, drawing on Castoriadis’ idea that imagination is not the opposite of reason but its generative core. If institutions are social-historical forms, then educators, as social agents, have the power, and the responsibility, to reimagine them otherwise.