Someone raises artificial intelligence. There is a brief, serious conversation about academic integrity. A working group is mentioned, possibly already formed, possibly yet to be convened. Someone notes that the IT department has been looking at detection tools. Someone else mentions that the academic regulations have been updated, or are being updated, or are under review. There is a general sense that the institution is taking the matter seriously.

And then the agenda moves on.

What is not discussed (what is almost never discussed at this level) is whether the curriculum itself is fit for the world graduates are entering. What is not on the risk register is the possibility that an entire cohort of students might complete a degree that was not designed for the professional environment they will inhabit. What is not being modelled is the reputational, competitive, and financial exposure that quietly compounds and accumulates when institutions mistake activity for strategy.

This piece is about that exposure. It is written for the leaders who are responsible for it, whether they know it yet or not.

Subscribe now

A Risk in Three Dimensions

The strategic risk of curriculum inertia in the face of AI is not a single threat. It arrives simultaneously from three directions, and the convergence of those pressures makes the current moment genuinely serious for institutional leadership.

Graduate employability and employer relations

Employers are not waiting for higher education to catch up. The World Economic Forum’s (2025) Future of Jobs Report, drawing on perspectives from over 1,000 leading global employers representing more than 14 million workers, projects that 39% of core skills required in the job market will change by 2030, with AI and big data literacy identified as the fastest-growing capability requirement across industries. This is not a distant forecast. It is a structural shift already reshaping what employers expect at the point of graduate recruitment.

The gap between institutional response and employer expectation is measurable and widening. Research by Portocarrero Ramos et al. (2025), examining AI skills and graduate employability across a sample of undergraduate and postgraduate completers, found that the majority of graduates reported feeling only “somewhat prepared” or “little prepared” to face the changes AI is driving in their labour markets, a finding the authors describe as revealing a “worrying gap between the evolution of the labor market and the actual preparation offered by educational environments” (p. 9). The Cengage Group (2025) Graduate Employability Report, surveying employers, graduates, and educators across the United States, found that while nearly 89% of educators believe their students are adequately prepared for the workforce, almost half of graduates (48%) report feeling unprepared to apply for entry-level positions, with 56% of those graduates citing job-specific skills, including AI fluency, as their primary deficit.

Across professional services, healthcare, engineering, media, and the public sector, AI fluency is moving rapidly from desirable to expected, not fluency in the sense of knowing how to use a particular tool, but the deeper capability to work alongside AI with critical independence: to interrogate its outputs, to understand its limitations, to make professional judgements that the technology cannot make.

When graduates arrive without this, it is noticed. What is changing is that it is increasingly being said out loud, in employer surveys, in graduate recruitment feedback, and in the conversations that senior leaders in industry have with vice-chancellors at dinners and on advisory boards. The gap between what degrees are producing and what employers now need is becoming a reputational issue, not just a pedagogical one. Institutions with strong employer partnerships have particular exposure here. The goodwill built over years of curriculum collaboration and placement development can erode quickly when the curriculum’s output no longer meets the standard on which the partnership was built.

Student recruitment and retention

The students entering higher education now have grown up with AI as a feature of their environment. They are arriving with expectations, sometimes inchoate, sometimes very explicit, about whether their degree will prepare them for a world they can already see forming around them.

The pace of student AI adoption is instructive here. The Higher Education Policy Institute’s (HEPI) Student Generative AI Survey 2026 (Stephenson & Armstrong, 2026), based on responses from 1,054 full-time UK undergraduates, found that 95% of students now report using AI in at least one aspect of their studies, with 94% using generative AI to support assessed work, figures that represent an extraordinary acceleration from just 66% and 53% respectively in 2024. The Digital Education Council’s (2024) global survey of 3,839 students across 16 countries found that 80% of respondents considered their institution’s current AI integration to fall short of their expectations. These are not students who are resistant to AI. They are students who use it daily and find that their institutions have not caught up.

The recruitment risk is not immediate, but it is real. It is the risk of a narrative forming, in student forums, in school common rooms, in the conversations that shape application decisions, that certain institutions, or certain types of institutions, are behind. Eighteen-year-olds making choices about where to study are increasingly asking not just what a university’s ranking is, but whether it is relevant. That is a new question, and it deserves a serious answer.

The retention risk is closer. Students who feel that their education is disconnected from the world they observe, who can see AI reshaping the professions they are training to enter while their curriculum proceeds as if nothing has changed, are likely to disengage. Completion rates, satisfaction scores, and graduate outcomes data move together and eventually appear in the metrics that matter to governing bodies and regulators.

Institutional reputation and rankings

The reputational dimension operates on a longer cycle but with greater permanence. The UK’s Quality Assurance Agency [QAA]’s guidance on generative AI (2024) signals that quality assurance expectations are moving, slowly but visibly, towards how institutions are redesigning learning and assessment for an AI-present environment, not merely how they are policing it. The updated UK Quality Code, now being implemented across the devolved nations of the UK, emphasises curriculum currency and the importance of preparing students to “succeed in their studies and progress their personal and professional lives”. A standard that, in the current environment, demands engagement with AI at the curriculum level.

More immediately, reputation in the sector is shaped by who is seen as leading and who is seen as managing. The institutions that are visibly, credibly engaging with AI at the curriculum level, not just the policy level, are beginning to differentiate themselves in ways that matter to the best academic staff, the most ambitious students, and the most valuable employer partners. The institutions that are not are quietly beginning to be categorised.

What Institutional Complicity Actually Looks Like

It is worth being precise about how this risk accumulates, because it does not look like negligence. It looks like reasonable, well-intentioned institutional behaviour.

It looks like a working group that produces a policy on AI use in assessment. It looks like a CPD session on detecting AI-generated text. It looks like a revised academic integrity framework. It looks like a technology procurement decision that brings AI tools into the student experience. All of these things are defensible. None of them addresses the curriculum.

The gap is structural. The mechanisms most institutions have for curriculum oversight, programme boards, academic boards, quality committees, periodic review, were not designed for the pace of change that AI represents. They operate on timescales measured in years. They are oriented towards compliance rather than strategic transformation. They are populated by people who are, individually, thoughtful and committed, but who are collectively operating within a system that is not built to ask the right questions at the right speed.

This is not a failure of individuals. It is a failure of architecture. And it is one that ultimately sits at the level of senior leadership, because only senior leadership has the authority and the overview to redesign it.

The Accumulation Problem

What makes this risk particularly difficult to manage is that it does not announce itself. There is no moment at which a dashboard turns red, no trigger event that makes the exposure visible. Instead, it accumulates.

Each cohort that graduates without AI-era curriculum preparation will, over time, see outcomes that reflect that gap. Each employer conversation that ends with polite concern rather than genuine confidence is a small erosion of a relationship that took years to build. Each prospective student who chooses an institution perceived as more forward-looking is a recruitment decision that doesn’t appear in any single dataset but nonetheless shapes the trajectory.

By the time the risk is legible in the metrics, in graduate employment rates, satisfaction scores, and application numbers, it will already be several years old. The decisions that created it will have been made, or not made, in leadership meetings that moved on to the next agenda item.

What This Moment Actually Offers

I want to be clear that this is not only a piece about risk. It is also a piece about opportunity, because the two are, in this case, inseparable.

The institutions that recognise what is at stake and act on it now are not simply avoiding a downside. They are positioning themselves for a genuine competitive advantage in a sector where differentiation is increasingly difficult to achieve. The ability to say, credibly and specifically, that your graduates are prepared for an AI-shaped professional world, and to evidence that claim through curriculum design, assessment practice, and employer feedback, is a proposition that will matter more with each passing year. As the World Economic Forum argues, embedding AI literacy, problem-solving, and critical thinking into every education programme is vital so that graduates know how to work with and complement the value being created by AI (Palsule & Leopold, 2025) .

This is a moment that rewards bold academic leadership. Not the boldness of grand announcements or technology procurement headlines, but the harder, less visible boldness of asking serious questions about what a degree is for, restructuring the governance mechanisms that answer that question, and investing in the faculty development that makes transformation possible rather than merely aspirational.

The institutions that do this will look very different in five years from those that don’t. The question for every DVC and PVC reading this is simple: which category is your institution currently on track to be in?

Next week, we will explore why current governance structures are preventing action and what it would take to redesign them.

This is the third in a series on AI and curriculum in higher education. The previous pieces on why most institutional responses miss the point and on what the curriculum needs to include are available here.

References

Cengage Group. (2025). 2025 Graduate Employability Report. Cengage Group. https://www.cengagegroup.com/news/press-releases/2025/cengage-group-2025-employability-report/

Digital Education Council. (2024). Digital Education Council Global AI Student Survey 2024 [DEC Global Survey]. Digital Education Council. https://www.digitaleducationcouncil.com/post/digital-education-council-global-ai-student-survey-2024

Palsule, H., & Leopold. (2025, December 15). Beyond the inflection point: The new forces shaping the transformation of work. World Economic Forum. https://www.weforum.org/stories/2025/12/work-transformation-skills-agility-growth/

Portocarrero Ramos, H. C., Cruz Caro, O., Sánchez Bardales, E., Quiñones Huatangari, L., Campos Trigoso, J. A., Maicelo Guevara, J. L., & Chávez Santos, R. (2025). Artificial intelligence skills and their impact on the employability of university graduates. Frontiers in Artificial Intelligence, 8. https://doi.org/10.3389/frai.2025.1629320

QAA. (2024). QAA advice and resources. https://www.qaa.ac.uk/sector-resources/generative-artificial-intelligence/qaa-advice-and-resources

Stephenson, R., & Armstrong, C. (2026). Student Generative Artificial Intelligence Survey 2026 (No. 199). Higher Education Policy Institute. https://www.hepi.ac.uk/reports/student-generative-ai-survey-2026/

World Economic Forum. (2025). The Future of Jobs Report 2025. World Economic Forum. https://www.weforum.org/publications/the-future-of-jobs-report-2025/

They join a mid-sized commercial firm. In their first week, they are handed a contract review task that would once have taken a junior associate two days. An AI tool completes a first-pass analysis in four minutes. Their supervising partner doesn’t ask them to redo it manually. They ask them to interrogate it: where might it be wrong? What has it missed? What are the ethical exposure points the model hasn’t flagged? What would they advise the client?

They have no framework for this. Their degree didn’t give them one.

This is not a hypothetical. Versions of this scene are playing out across law firms, hospitals, design studios, newsrooms, engineering consultancies, and classrooms right now. And it points to something curriculum designers can no longer defer: the question of what, exactly, graduates need to be able to ‘do’ in a world where AI is a permanent professional collaborator.

Last week, I argued that higher education has been focused on the wrong emergency, obsessing over plagiarism and assessment integrity while the deeper curriculum crisis quietly compounds. This week, I want to be specific about what the right response looks like.

A Framework for AI-Ready Curriculum

Before listing competencies, it is worth naming the design principle that should underpin them all.

The goal is not to teach students about AI. It is to develop graduates who can think and work ‘alongside’ AI with critical independence; who can leverage its capabilities without being captured by them, and who understand enough about how these systems work to know when to trust them, when to question them, and when to refuse them entirely. The distinction between teaching about AI versus learning to work alongside it is crucial. A module on “AI literacy” that surveys the current landscape of tools and explains how large language models work provides a useful background. It is not, on its own, a curriculum transformation. What follows are the competencies that actually are.

The Core Competencies

Critical Evaluation of AI Outputs

This is the foundational skill, and it is more demanding than it sounds. It is not simply about fact-checking. It is about developing the disciplinary judgment to assess whether an AI output is not just accurate but also appropriate, whether it has framed the problem correctly, surfaced the right considerations, applied sound reasoning, and produced something fit for the specific professional and ethical context at hand.

In medicine, this means a student who can assess whether an AI diagnostic suggestion has appropriately weighted the patient’s symptoms. In journalism, it means a reporter who can identify not just factual errors but the subtle framings and omissions that a generative tool might introduce. In architecture, it means a designer who can evaluate whether an AI-generated scheme meets not just technical requirements but human ones. This competency cannot be taught in a generic way. It must be embedded in disciplinary practice, which is precisely why it requires curriculum redesign rather than a standalone module.

One of the things I have always enjoyed in marking assignments has been discovering an insight, unique to the author, drawn from some tangential source in an apparently unrelated discipline. Relying exclusively on AI forecloses exactly this kind of serendipity; it locks the learner into the model’s own paradigm and epistemological framework. AI is not going to speculate about the relevance of an alternative perspective; it was never trained to surface it.

Prompt Literacy and Problem Decomposition

Working effectively with AI tools requires a cognitive skill that higher education has not historically needed to name: the ability to decompose a complex problem into well-formed, productive queries. This is, at its core, a thinking skill; it demands clarity about what you’re actually trying to find out, what constraints apply, and what a good answer would look like before you begin. Students who struggle to articulate their thinking will have difficulty directing AI effectively. Conversely, developing strong prompt literacy can serve as a powerful diagnostic for deeper analytical capability. Curricula that build this skill are, in effect, building critical thinking under a new name.

In my day as an undergraduate (1985–88), this meant taking an essay question, identifying the core knowledge it sought, and then spending hours in the library stacks, not just finding sources, but stumbling across unexpected ones, following tangential threads, discovering connections that no search could have predicted. It was an education in intellectual serendipity. That particular privilege has been quietly removed from most university libraries, and AI accelerates its disappearance. The question for curriculum designers is how to deliberately restore it.

Ethical Reasoning in AI-Mediated Contexts

This goes well beyond a single ethics module. Every discipline now faces AI-specific ethical terrain: questions of bias in automated decision-making, accountability when AI contributes to professional judgements, transparency with clients and stakeholders about AI use, data privacy in AI-assisted workflows, and the environmental costs of large-scale model use. These are not abstract philosophical puzzles. They are live, professional questions that graduates will face, often quickly and with little institutional support.

Ethical reasoning here also includes something more prosaic but important: knowing when AI is simply the wrong tool for the task. I recently found myself discussing with a political scientist how to establish an efficient leafletting approach in a ward ahead of a local election. After working through how complex the prompt would need to be, and reflecting on the ethics of aggregating multiple public data sources, we concluded this was a ‘sledgehammer to crack a nut’, and reverted to printing out a Google Maps view. Sometimes the most sophisticated judgement is recognising that AI adds cost, complexity, and risk where a simpler approach would serve perfectly well.

Human Judgement and the Irreducibly Human

One of the most important things a curriculum can do right now is help students identify what AI cannot do well and develop confidence in those domains. Empathy, relational intelligence, creative originality, moral courage, contextual wisdom, the ability to hold ambiguity and make judgements under uncertainty: these are not soft skills. They are human skills. In an AI-augmented world, they are the differentiating ones.

This is partly a curriculum design question and partly a pedagogical one. Assessment that only tests knowledge retrieval, the kind that AI can now perform with ease, is not just vulnerable to plagiarism. It is educationally inadequate. Designing for the irreducibly human means assessing students on what machines cannot replicate: judgment, creation, argumentation, and reflection.

For those familiar with my work, you will know that I am avowedly committed to extending the use of learning outcomes beyond Bloom’s ubiquitous cognitive domain to encompass five domains of learning: the Affective (values), the Metacognitive (self-reflection and learning to learn), the Psychomotor (practical skills, including how to use AI and software), and the rich, complex array of very human skills that fall under the Interpersonal domain. In the age of AI, the cognitive domain has arguably become the least important for faculty to teach in isolation. Any programme or course that does not attend to all five domains is, I would argue, failing its students.

Adaptability as a Curriculum Outcome

Perhaps the most honest thing we can say about AI and the future of work is that we don’t know exactly what it will look like in five or ten years. The tools available to a graduate in 2030 will be substantially different from those available today. This means that one of the most important things a curriculum can deliver is not knowledge of current tools but the capacity to adapt to tools that don’t yet exist. This is a metacognitive competency: learning how to learn in rapidly shifting technological environments, developing comfort with uncertainty, and building the habit of critically engaging with new tools rather than passively adopting them. It is, in many ways, the oldest educational virtue reframed for a new context.

What This Requires of Curriculum Design

These competencies do not slot neatly into most existing module or course structures. They require a different approach to curriculum architecture, built around a few clear principles.

Integration over addition. The temptation is to respond to AI by adding things: a new module here, a workshop there. But competencies like critical evaluation and ethical reasoning cannot be delivered in isolation. They need to be woven through the curriculum, modelled in disciplinary practice, and assessed in context. Adding a standalone “AI skills” unit while leaving the rest of the curriculum unchanged is the curricular equivalent of rearranging deck chairs.

Assessment redesign is not optional. If the assessments haven’t changed, the curriculum hasn’t changed. Assessment drives learning (whether we like it or not), and an assessment that a capable AI prompt can complete is no longer doing its job. This doesn’t mean banning AI from assessments; in many cases, the opposite is true. It means designing tasks that require students to demonstrate what they and they alone can contribute: their reasoning, their judgement, their voice, their ethical stance.

Curriculum as a living document. In most institutions, curriculum review is a periodic, laborious process that happens on a five or seven-year cycle. That cycle is not compatible with the pace of AI development. Programmes need mechanisms for more agile review, not wholesale redesign every year, but the ability to update examples, adjust assessments, and incorporate emerging practice without waiting for the next formal review window.

The Faculty Development Corollary

I argued last week that mandatory, structured CPD for all faculty is a prerequisite for any of this to work. This week’s proposed approach makes that case even more concretely.

A faculty member cannot design assessments that develop critical evaluation of AI outputs if they have not developed that capacity themselves. They cannot embed ethical reasoning about AI in their teaching if they have not seriously engaged with the ethical questions their own discipline faces. They cannot model adaptability if they are not modelling it.

Professional development here is not about training faculty to use AI tools. It is about creating the conditions for genuine pedagogical imagination, the kind of thinking that produces curriculum change, not just policy compliance.

It would be easy to read all of this as a burden, another set of demands on already stretched institutions and faculty. I want to end by naming the other side of that.

AI is forcing higher education to ask questions it should have been asking for decades. What is a degree actually for? What can only a human do, and how do we cultivate it? What does it mean to prepare someone not just for their first job but for a career that will span technological shifts we cannot yet see? These are not new questions. They are the oldest questions in education. AI has simply made them urgent in a way that is hard to ignore.

The institutions that treat this moment as an opportunity to do that thinking seriously, rather than a crisis to be managed defensively, will produce graduates who are genuinely prepared for what’s coming. And they will deserve the reputation for it.

This is the second in a short series on AI and curriculum in higher education. The first piece, on why most institutional responses are focused on the wrong problem, is available here.

Next Week - The Cost of Inaction: an Institutional Risk. We will explore reputational and competitive risk, reframing curriculum inertia not as a pedagogical failing but as a strategic liability.

Subscribe now

Leave a comment

This Substack is Free: Buy me a coffee?

This is where much of higher education finds itself right now with artificial intelligence.

What Institutions Are Actually Doing

Walk into most universities and ask what they’ve done in response to AI, and you’ll get a familiar answer. They’ve updated their academic integrity policies. They’ve invested in detection tools (many of which are unreliable and most of which are already being outpaced). They’ve debated endlessly about what constitutes AI-assisted plagiarism and how to prove it. Some have gone further, banning the use of AI in assessments entirely, as if the goal were to preserve the examination hall as a kind of heritage site.

At the other end of the spectrum, institutions have made significant investments in AI-integrated management systems, streamlining admissions, student services, timetabling, and reporting. This is, in its own way, rational. Administrative efficiency has genuine value.

But here is what is striking about both responses: they are almost entirely about the institution’s needs rather than the learner’s future.

The plagiarism panic is fundamentally about protecting the integrity of existing assessment systems. The administrative AI deployment is about operational efficiency. Neither asks the harder question: What does a person need to know, and be able to do, in a world reshaped by AI, and is our curriculum preparing them for it?

The Misalignment at the Heart of the Response

There is a useful distinction between first-order and second-order problems.

The first-order problem is the one that lands on your desk and demands immediate attention: a student submits work that may have been AI-generated. What do you do? This is urgent, concrete, and politically uncomfortable. Of course, institutions respond to it.

The second-order problem is slower, quieter, and far more consequential: are we designing learning experiences that equip graduates for a world in which AI is a permanent, pervasive feature of professional and civic life? This question doesn’t send an email. It doesn’t trigger a disciplinary hearing. It simply accumulates, silently, as cohort after cohort graduates underprepared.

Higher education has historically been good at responding to first-order problems and poor at anticipating second-order ones. The AI moment is severely testing that weakness.

The risk is not that students will use AI to cheat on an essay. The risk is that we will spend five years perfecting our response to that problem while failing to redesign curricula that are fit for the world those students are entering.

The Faculty Development Gap Nobody Is Talking About

There is a second failure running alongside the curriculum one, and it may be even more foundational: most faculty are being asked to navigate a profound pedagogical shift without any structured support. There are clearly exceptions, and if you are one of them, these statements may seem unreasonable. But explore the corridors of your own campus and see how many of your colleagues are truly like-minded.

Consider how other professions handle this. Lawyers, doctors, accountants, engineers, nurses; virtually every regulated profession requires practitioners to complete a minimum number of hours of continuing professional development each year to maintain their licence to practise. CPD is not optional. It is not left to individual motivation. It is a structural guarantee that professional knowledge does not stagnate.

Higher education has no equivalent. A professor appointed in 2005 is under no institutional obligation to update their understanding of how learning, technology, or their discipline has changed since then. This was always a quiet weakness in the system. In the current moment, it is an acute one.

You cannot ask faculty to redesign the curriculum for an AI-shaped world if they have not had the opportunity, with support, to develop a sophisticated, up-to-date understanding of what that world looks like. How AI is actually being used in their disciplines, what it can and cannot do, where it is changing the nature of work and knowledge production, and what that means for what students should learn and how.

Good intentions are not enough. Individual enthusiasm is not enough. What is missing is a structural commitment: mandatory, resourced, ongoing professional development for all faculty, not a one-off workshop on spotting ChatGPT, but sustained engagement with the questions that AI is forcing onto the table.

This is not a radical idea. It is standard practice in almost every other profession. The question is why higher education has so long believed itself exempt from it.

What “Missing the Point” Actually Costs

It is worth being clear about the stakes, because the consequences of curriculum inertia are not abstract.

Graduates are entering labour markets in which AI fluency, not just familiarity, but genuine, critical, adaptive fluency, is increasingly a baseline expectation. They are entering professions that are being restructured around these tools. They are becoming citizens who will need to navigate AI-mediated information environments, automated decision-making, and the ethical and political questions that accompany them.

If the curriculum they studied was designed around concerns that predate this moment, or worse, designed defensively ‘against’ AI rather than thoughtfully ‘for’ a world that contains it, they are being sent out underprepared. And they will know it.

The institutions that will serve their students best in the next decade are not the ones that locked down their assessment policies most effectively in 2025. They are the ones who asked the harder, slower, more important question: What does learning need to become?

A Provocation, Not a Conclusion

This piece is deliberately diagnostic. The case for ‘what’ curriculum should contain in response to AI, the specific competencies, dispositions, and design principles that should be shaping what we teach, is a conversation for next week.

But the diagnosis matters because you cannot design the right solution if you have misidentified the problem.

Higher education is not facing an assessment crisis. It is facing a curriculum crisis, compounded by a professional development crisis. The institutions that recognise this and act on it will look very different from those that spent the decade perfecting their plagiarism detection.

The question is which kind of institution you want to be.

Subscribe now

Leave a comment

Next week: What AI actually demands of curriculum: the competencies, principles, and design choices that should be shaping what we teach next.

The pattern is remarkably consistent. A review is commissioned, usually in response to external pressure, a quality audit, an accreditation requirement, or a new institutional strategic plan. A committee is formed, typically drawn from senior academics and professional staff. Consultation happens, surveys, focus groups, and meetings with student representatives. A report is produced with recommendations. Leadership accepts the recommendations. An implementation plan is written. Everyone returns to their normal workload.

Eighteen months later, the curriculum looks almost identical to what it was before the review began. The documentation has changed. The practice has not.

Why does this happen so reliably? The answer lies in what curriculum reviews are actually designed to do. Most reviews are fundamentally governance processes. They are built to satisfy external stakeholders, quality agencies, professional bodies, and institutional governing boards. Something systematic has been done. The criteria they use to evaluate curriculum are almost always structural: programme architecture, credit frameworks, graduate attribute mapping, and alignment with qualification standards. These are not unimportant things. But they are not the things that determine whether students actually learn well.

What determines that is learning design, the decisions made about how content is sequenced, how learning activities develop understanding over time, how assessment genuinely measures what it claims to measure, and how feedback creates forward momentum rather than simply recording a grade. Curriculum reviews almost never examine learning design in any depth. They examine what is being taught. They say almost nothing about how.

This produces a characteristic outcome: a curriculum that has been restructured without being rethought. New programme titles, revised credit weightings, updated graduate attributes, and underneath it all, the same lectures, the same assessments, the same implicit assumptions about how students learn. The form changes. The substance does not.

There is a second problem that is less often explored. Curriculum reviews concentrate their energy at the point of deliberation, the committee meetings, the consultations, the drafting of recommendations, and then assume that implementation will follow naturally from the authority of the report. It rarely does. Academic staff who were not meaningfully involved in the review have no particular reason to change their practice in response to it. A recommendation that programmes should embed authentic assessment more consistently is not, by itself, a reason for a lecturer who has been running the same assessment process for fifteen years to do anything differently. Change in academic practice requires development, not instruction.

The institutions that break this pattern have recognised something important: curriculum review is most effective when it is designed as a professional development process, not a governance one. This does not mean abandoning governance functions, external accountability matters, or documentation; documentation has its place. It means building the review so that it develops capability as it proceeds, rather than assuming capability will materialise once the report has been accepted.

In practice, this looks like a review process in which academic staff are not merely consulted but genuinely engaged in examining their own learning design decisions. It looks like workshops embedded in the review timeline, not bolt-on additions, but integral stages where the evidence gathered in consultation becomes the basis for structured professional conversation. It looks like review criteria that include questions about pedagogy alongside questions about structure: not just ‘does this programme map to the graduate attributes?’ but ‘does the assessment in this programme actually develop the capabilities it claims to assess?’

This kind of review takes longer and requires more from institutional leadership. It requires a DVC, PVC or Dean who is prepared to treat the review as an investment in long-term capability, not a compliance exercise to be completed. It requires academic developers who are involved from the outset, not brought in at the implementation stage, to run a workshop on the new framework.

But it produces something that conventional reviews almost never do: academic staff who understand why the curriculum has changed and who have developed the design capability to make that change real in their teaching.

Documents do not change curricula. People do. And people change when they have been genuinely developed, not merely informed.

If this pattern resonates with something you are navigating in your own institution, I would be glad to hear about it. Comment please, I read every response.

Leave a comment

Next issue: What AI actually demands of curriculum design — and why most institutional responses are missing the point.

—

Dr Simon Paul Atkinson (PFHEA)

Learning Design & Curriculum Strategy Consultant

sijen.com | Capable Institutions on Substack