SRHE Blog

The Society for Research into Higher Education

Monthly Archives: August 2026


by Leave a comment

Vibe-coding and co-creation: human-AI collaboration in higher education

by Andrew Williams

The debate around artificial intelligence (AI) in higher education has largely focused on students. Are they using ChatGPT in assessments? How do we maintain academic integrity? How should teachers and curricula adapt?

Much less attention has been given to academics working with generative AI (GenAI) as a creative collaborator or co-pilot, rather than merely attempting to mitigate its impact. The potential for productive and effective AI collaboration deserves greater consideration because it challenges longstanding assumptions about expertise, creativity and educational design.

Ever since the release of ChatGPT in November 2022, I have been fascinated by how intuitive the interaction with GenAI tools can be and the ease with which ideas can be explored through ordinary conversation. GenAI is a sophisticated technology that allows users to test possibilities, refine outputs, and rapidly turn preliminary concepts into usable material.

My background is in biological sciences, with no grounding or expertise in computer science, software engineering or coding. For me, this technology has helped me overcome technical barriers and allowed me to explore new ways to teach and to develop bespoke educational resources. I envisioned a learning experience for my undergraduate medical science students (studying human genetics and evolution) that moved beyond static textbook diagrams and dry explanations to a more interactive and visually stimulating format. Realising that vision required a simulation that let students tweak parameters, observe real‑time population dynamics, and grasp complex evolutionary ideas.

Traditionally, such a project would rely on grants, external software developers, or months of self‑taught coding. For most of us, the gap between an innovative pedagogical idea and a digital reality was confined by technical expertise. Today, through iterative collaboration with GenAI, natural language can function as a design interface, allowing educators to transform ideas into working educational tools.

I was also intrigued by the ever improving, and often publicised, capabilities of GenAI tools as coding agents. With no coding experience myself, I decided to try my hand at ‘vibe-coding’.

Natural language as a design interface

‘Vibe-coding’ is an emerging term within technology communities that describes a process where functional intent is communicated via natural language prompts, allowing AI to generate and refine code. This allows the user to focus on what the tool should do rather than how it must be coded. Essentially, ‘vibe-coding’ involves directing the AI to write and co-develop code aligned with your vision and pedagogical objectives. In other words – no code, no problem!

My recent research describes an iterative prompting strategy and the outcomes of a human-AI collaboration to create an interactive simulation that models certain aspects of evolution (Williams, 2026a). Instead of learning JavaScript or working with software developers, I co-created an application with ChatGPT using only natural language prompting. The result was a browser-based simulation capable of supporting inquiry-based learning and formative assessment in undergraduate education.

Rather than blindly generating content based on loosely defined instructions, the human-AI collaboration involved a structured iterative prompting strategy, careful testing of AI-generated output and re-prompting until the desired aims of the project were met. This enabled the creation of an interactive simulation in which students could investigate biological systems, manipulate variables, export datasets and construct evidence-based explanations (figure 1). This allowed students to move beyond consuming information, to generating and interpreting data themselves in real-time.

Figure 1. Snapshot of ‘Survival Island’ – an interactive simulation.

The Illusion of automated expertise

There is a prevailing narrative that AI democratises expertise, in that it allows anyone to do anything more quickly and with less effort. My experience of ‘vibe-coding’ suggested something different. The  credibility of the final simulation depended on actively embedding disciplinary expertise directly into the co-design process.

While the AI had removed the technical barrier (the coding), it had increased the importance of my academic expertise and oversight. The AI could generate a thousand lines of code in seconds, but it couldn’t tell me if the code represented a plausible scientific simulation. My epistemic judgement was required to verify the AI-generated output. The final student-ready simulation required multiple rounds of iteration, testing and re-prompting, until the collaboration achieved a credible and biologically grounded application suitable for teaching.

Rather than replacing my expertise, ChatGPT provided capabilities I do not possess: writing HTML; JavaScript; debugging code. I contributed disciplinary knowledge, pedagogical intent and continuous evaluation. Every feature reflected educational decisions rather than technical ones. The simulation succeeded because the educational need was clearly defined before AI generated the code.

Higher education has often treated GenAI as an automated assistant or as a problem to be mitigated. Increasingly, it may be more productive to think of AI as a collaborator whose contributions depend upon human judgement. Expertise remains indispensable, as humans ultimately remain responsible for judging whether outputs are educationally, scientifically and ethically sound.

Beyond the prompt: the AI literacy gap

Universities have understandably invested considerable effort in developing student AI literacy. This project led me to a broader appreciation of the need to promote teacher AI literacy across higher education, a critical area often overlooked.

AI literacy for teachers is not just about knowing how to write a good prompt or choosing the best online GenAI tool. Rather, it is a multi-dimensional competency that considers foundational knowledge (prompting, tool use), ethical, emotional and technical proficiencies. For educators to co-create teaching resources in collaboration with AI, for example through vibe-coding, more than technical curiosity is required.

The iterative ‘vibe-coding’ process I followed mirrors the ‘critical evaluation’, ‘epistemic’ and ‘innovation’ dimensions of an integrated AI literacy framework for higher education teachers (Williams, 2026b). This recently published conceptual framework comprises eight core AI literacy dimensions, which you can access in full using the link.

In the current context, the important  AI literacy dimensions include:

  • Critical evaluation: the ability to treat the AI model as a “black box” and rigorously verify outputs against authoritative sources.
  • Epistemic authority: a clear understanding of the probabilistic nature of GenAI and the domain-specific expertise to judge AI output, thereby retaining human agency during interactions with AI.
  • Innovation mindset: the willingness to move from being a consumer of tools to a co-creator of bespoke learning environments, including an understanding of how to collaborate effectively with GenAI.

These are becoming essential academic capabilities rather than optional technical skills. This transition presents considerable challenges, as academic practice has historically been predicated on either individual resource creation or the outsourcing of technical development to third parties. GenAI now enables academics to become directors of increasingly sophisticated creative processes. They define the educational challenge, AI performs the technical implementation, and academics evaluate and verify the final outcome.

As technical production becomes easier with GenAI, educational judgement becomes increasingly important. Our value is no longer found in the technical labour of production, but in our ability to define the educational challenge, critique the iterations, and curate the final outcome. The “vibe” we are coding is our pedagogical intent.

Key takeaways

  • AI is a tool, not a replacement.
  • Human-AI collaboration enables pedagogical creativity.
  • ‘Vibe-coding’ uses AI’s coding capabilities but demands epistemic oversight.
  • Teacher AI literacy is multi‑dimensional.
  • Institutional support is required to build that AI literacy.

GenAI is redefining the landscape of higher education, not through incremental improvements to traditional teaching materials, but by enabling entirely new educational content to be created. However, despite the promises of this new technological revolution, we still need robust evidence about student learning gains, engagement, accessibility and the sustainability of AI-generated educational resources.

It is easy to provide AI tools for HE educators to use. The challenge is to give educators the technical, critical and ethical literacy required to use them effectively.

Andrew Williams is a Professor (Teaching) in the Faculty of Medical Sciences at UCL. He teaches immunology and cell and molecular biology to undergraduate and postgraduate students. His research aims to explore the impact of artificial intelligence (AI) on assessment and learning in higher education and the ways in which we can promote student and staff AI literacy.


by 1 Comment

The quiet epiphany: what ‘Stoner’ teaches us about academic life

by Dr Goran Erfani

Over the recent holidays, I finally had a chance to read Stoner by John Williams. This novel has quietly become a cult classic since its original publication in 1965, despite selling almost nothing at the time. It tells the story of William Stoner, who teaches English literature at the University of Missouri (US) for forty years, and whose life, on the surface, is unremarkable: no great triumphs, no dramatic falls, just the long accumulation of an academic career.

And there is as yet very little about it that feels outdated. The perceived career barriers, the departmental politics, the strange intimacy of a life built around a library and a lecture hall all read as though they were written last year, not eighty years ago. Reading the novel raised a few academic-life questions for me: how much of a career is really chosen versus tripped into, how little control academics may have over the wider forces that shape their work, why the office can so easily become a refuge from everything else, and why recognition so often arrives too late to matter.

The epiphany that redirects a life

The protagonist, Stoner, begins his journey as an agriculture student, sent to university by his farming parents to learn better methods for the land he’s expected to inherit. Then, almost by accident, he takes an optional literature module which was not very popular among students from other departments. A single class, a moment his teacher, Archer Sloane, later calls his epiphany, reroutes the entire course of his life. He abandons the farm, and the future his parents assumed for him, to become an academic instead.

This story is a powerful reminder of something many of us recognise from our own paths: academic careers are rarely the exclusive product of a grand plan. More often our careers turn on one unplanned module or a seminar, one conversation, one person who saw something in us before we saw it ourselves. The novel is generous in this sense, and it doesn’t treat Stoner’s choice as either triumphant or tragic, just as the quiet hinge on which everything else swings.

When the wider world intervenes

The novel doesn’t pretend the wider world stays outside academia, although Stoner’s career unfolds almost entirely inside university walls. His two closest friends from graduate school, Dave Masters and Gordon Finch, both go off to fight in the First World War: Masters is killed within months, while Finch survives and later becomes dean of the faculty. Despite the prevailing social disfavour, Stoner stays behind, choosing to finish his PhD rather than enlist. The novel doesn’t treat this decision as cowardice, or even really as a deliberate choice. In fact, Stoner just quietly stays where he was, against a backdrop of events entirely beyond his control. He rarely dwells on it afterwards, but the war reshapes his career as much as his friendships: with so many young men gone to fight, teaching posts open up that might otherwise have taken him years to reach. He ends up moving up faster than he probably would have otherwise, and it’s not something he ever really has to reckon with. The same war that costs his closest friend his life is the one that clears space for Stoner’s.

It’s a useful corrective to how we sometimes talk about academic careers as though they were shaped only by our ambition, ability, and choice. Wars, recessions, pandemics, policy changes, and funding cuts imposed from far outside a department are just as often what redirect them. And as Stoner’s case shows, that redirection isn’t always a loss, since it sometimes opens a door for one person precisely because it closes one for somebody else. The university, however insulated it can feel, has never really been separate from the world beyond it.

Perfunctory interest, and the retreat into work

What struck me most, though, was how Williams describes Stoner’s relationship to everything outside his academic work. His marriage fails early and never really recovers; his daughter drifts from him; and increasingly he treats his own home life with a kind of ‘perfunctory interest’— present, dutiful, but never fully engaged. The office becomes his real home. Long hours grading papers or losing himself in medieval literature aren’t just work; they’re an escape from a domestic life he can’t fix and, eventually, stops trying to.

It’s an uncomfortable thing to recognise, but many academics will know some version of it: the pull of the office as sanctuary, ‘one more hour of reading’ as a fairly acceptable form of avoidance. The pressure to keep publishing more and more doesn’t help either. I’ve caught myself judging a whole day by how much I’d written, not by much else that happened in it. The novel doesn’t moralise about any of this. It just leaves you sitting there with it.

A career recognised too late

Stoner’s professional life follows its own quiet cruelty. A departmental rival, Lomax, pushes him away from the teaching and supervision he loves most and is best at. Recognition, becoming a full professor in more than title, comes only at the very end of his career, at the point of retirement, when it can no longer change anything. It’s a sharp, understated commentary on how academic institutions can delay or withhold recognition until it’s almost beside the point.

This is also a familiar pattern in academic life more broadly. Promotion panels and tenure timelines are often slow to catch up with what someone has actually contributed, and by the time recognition arrives, sometimes years after the work that earned it, the person it’s meant for has usually moved past caring about it.

The balance we still struggle with

Reading it today, it’s hard not to measure how far academic culture has moved on this front, or how little. It also raises a potentially uncomfortable question: would Stoner survive in today’s academic culture? Everything quiet about him, the private choices, the unhurried devotion to reading and writing, is perhaps what current academic life has the least patience for. His love of the discipline was rooted in learning and literature for their own sake, not in being judged primarily by his publication outputs — an orientation that sits awkwardly with how academic worth is now measured by metrics, the REF (Research Excellence Framework), or grant income.

But the main loss may not be the metrics themselves so much as what they’ve gradually displaced: unstructured time to simply read, think, or sit with a text without needing to turn it into a deliverable output. Stoner’s long hours in the library produced nothing that would count on a REF return, and that, more than any single policy, is what makes his kind of academic life hard to imagine now.

Even so, higher education institutions (in the UK and beyond) now talk openly about work-life balance, wellbeing, and boundaries in a way Stoner’s generation never could, and most of us have absorbed at least the language of it (I catch myself using it in meetings I attend). Yet the pull he feels towards the office, the sense that there’s always something else waiting to grade, read, or write, will still be familiar to plenty of academics negotiating where work ends and home and social life begin. We have the vocabulary now to name the imbalance Stoner never thought to question. I’m not sure we’re any better at actually living with it.

Why it still resonates

Stoner isn’t a true story, but it captures something true about academic life all the same: the ongoing tension between institutional demands and personal passion, the way one small decision can redirect a life, and the cost of retreating into work instead of confronting what’s difficult at home. It’s a quietly unsentimental book. It doesn’t romanticise academic life or condemn it. It simply shows it in full and trusts the reader to feel the weight of it. Reading it now, in a very different era of higher education, I find it less a period piece than a mirror.

Reference: Williams, J (1965) Stoner. New York: Viking Press. Currently available through NYRB Classics (US) and Vintage Classics (UK).

Dr Goran Erfani, FHEA, is a Senior Research Fellow in the School of Healthcare and Nursing Sciences at Northumbria University, Newcastle upon Tyne. His research interests include people-place-health dynamics, public health, digital inclusion, and health equity. Email: goran.erfani@northumbria.ac.uk. X/Twitter: @GoranErfani.


by Leave a comment

Learning analytics won’t close awarding gaps. Better teaching will.

by David Mather

Universities have invested heavily in understanding awarding gaps. We know more than ever about differential outcomes between student groups. We have sophisticated learning analytics platforms, increasingly accurate predictive models and growing confidence in our ability to identify students who may disengage, withdraw or underperform. This is important work, as it has helped institutions move away from reactive support towards earlier and more informed intervention.

Despite this, awarding gaps remain stubbornly persistent. Perhaps that is because we have spent the last decade using data to identify students rather than improve teaching. If we have become better at identifying students who are likely to struggle, why have we made less progress in changing the outcomes we are trying to improve?

Perhaps the answer lies in the purpose we have given our data. Learning analytics has become one of higher education’s greatest success stories. Universities now know more about their students than at any point in the sector’s history. Attendance, engagement with virtual learning environments, assessment patterns and a range of demographic indicators provide increasingly sophisticated pictures of student behaviour (Romero & Ventura, 2024). Most institutions now use this information to answer a single question: who is most likely to struggle?

It is an important question, but I am no longer convinced it is the most educational one.

Learning analytics has become remarkably good at supporting institutional decision making. It helps us identify educational risk, target student support and improve continuation. However, prediction is not pedagogy. Identifying students who may struggle is not the same as understanding how teaching might change their trajectory. Awarding gaps are not created by dashboards. They emerge through complex interactions between curriculum, assessment, expectations, recognition, belonging and students’ experiences of learning. They will not be closed by increasingly sophisticated prediction alone. Therefore, the assertion of this article is simple: better teaching closes awarding gaps. The role of data should be to improve teaching rather than improve prediction.

The distinction that I have presented became clear to me long before I joined higher education. My career began in further education, where I taught and led in sixth form colleges. Like many colleagues working in that sector, I became accustomed to using ALPS, a value added system that compares students’ outcomes with those expected from their prior attainment (ALPS Education, 2025). The numbers themselves were rarely the most interesting part of the conversation – the pedagogic conversations they created were.

Course teams did not gather simply to identify students who might miss their target grades. They asked why one assessment had produced stronger outcomes than another. They debated whether curriculum sequencing was helping students build confidence. They compared approaches to feedback and questioned whether different teaching strategies had influenced progress. Data became a catalyst for professional dialogue. It encouraged teachers to think critically about pedagogy rather than simply monitor performance.

One observation emerged repeatedly. Students at either end of the attainment spectrum were visible. Those achieving well continued to receive stretch and challenge. Those facing significant barriers rightly attracted additional support. However, those in the middle of those two poles often received much less attention. These were students who attended lessons, completed assignments and achieved respectable grades. They were progressing, but they had begun to plateau. Their potential remained largely hidden because nothing within the data suggested they required intervention.  I have become increasingly convinced that higher education contains the same group.

Earlier this year I described those between the two poles as the silent middle (Mather, 2026a). My argument then focused on belonging. Universities have become increasingly effective at identifying students who excel and students who experience significant challenge. Those in between often remain unnoticed because they continue to progress through their programmes without attracting concern. Progress, however, should never be mistaken for fulfilled potential. My earlier articles have led me to a conclusion that the silent middle is not simply a group of overlooked students – they reveal something more fundamental about the questions we ask of our data.

Learning analytics tells us who is likely to struggle. It tells us much less about where teaching is most likely to make the greatest educational difference. Those are not the same question. One helps universities manage institutional risk. The other has the potential to improve educational practice. The distinction is significant because universities cannot close awarding gaps through prediction alone. They will close them by improving teaching, strengthening relationships and creating learning environments in which more students can succeed. Learning analytics has an important role to play in that ambition, but only if we stop asking it to tell us merely who needs support and start asking what it can teach us about teaching itself. The question is not whether universities should continue to invest in learning analytics (they should). The question is whether we have become too comfortable allowing learning analytics to define the educational problem.

For the past decade, learning analytics has largely been framed as an exercise in prediction. Institutions seek to identify students who may withdraw, fail or disengage so that support can be targeted before those outcomes become inevitable (Romero & Ventura, 2024). This represents an advance over retrospective approaches that identified concerns only after students had already fallen behind. Early identification and timely-bound support are crucial. The sector should not lose either.

However, prediction is not an educational strategy. Prediction tells us which students are most likely to experience difficulty. It tells us very little about what educators should do differently. A dashboard cannot redesign a curriculum, improve formative feedback, strengthen relationships between students and staff or create cultivate belonging. All of those are fundamentally pedagogical acts.

Awarding gaps emerge through pedagogy as much as they emerge through student characteristics. Curriculum design, assessment practices, expectations, feedback, representation and relationships all influence whether students flourish or merely progress. All of this means that learning analytics should be judged not simply by its ability to predict outcomes, but by its ability to improve teaching.

The distinction between prediction and teaching quality has profound implications for educational leadership. Every university leader understands that resources are finite. Personal tutors, learning developers, academic skills teams and lecturers cannot provide intensive support to every student throughout the academic year. Decisions therefore have to be made about where educational effort should be concentrated. Those decisions are often informed by indicators of risk. Students who stop attending, miss assessments or disengage understandably receive attention. That is appropriate and necessary, but I wonder whether we have overlooked another question: where is teaching most likely to change a student’s trajectory?

That question shifts the purpose of learning analytics completely. Instead of asking who is most likely to fail, we begin asking where educational opportunity exists. Instead of identifying institutional risk, we begin identifying pedagogical opportunity. Those are not competing priorities – they are different ways of thinking about educational gain.

My experience in further education suggests that this distinction is crucial. ALPS data did not conclude conversations about attainment – it started them. Teachers debated why one class had responded differently from another, why particular assessment approaches appeared more effective and which students seemed to benefit most from timely feedback or increased challenge. Data did not replace professional judgement – it sharpened it. Higher education has an opportunity to claim that mindset.

Freire (1970/2005) argued that education is transformative because it encourages dialogue, reflection and action rather than passive acceptance. His critique was directed at pedagogy, yet it also offers a useful way of thinking about educational data. Learning analytics should not become another form of educational banking in which institutions accumulate information without fundamentally changing practice. Data has educational value only when it changes the conversations educators have about teaching.

Addressing data for the purpose of informing pedagogy and practice should intersect with the conversation about awarding gaps. Universities have invested considerable effort in understanding differential outcomes across student groups. That work has transformed institutional awareness of structural inequalities and rightly challenged assumptions that degree classifications simply reflect individual effort or ability. Research has consistently demonstrated that awarding gaps emerge through complex interactions between curriculum, assessment, belonging, institutional culture and wider social structures rather than any single deficit within students themselves (Mountford-Zimdars et al, 2015). Addressing those gaps therefore requires sustained institutional commitment.

The institutional commitment I refer to is where the silent middle re-enters the conversation. Those in the silent middle are rarely identified as requiring support because they continue to engage, submit work and progress through their programmes (Mather, 2026a). They do not generate alerts because they are doing precisely what institutional systems expect them to do. Many achieve marks in the middle classifications throughout their degrees. They are neither failing nor flourishing – they simply continue. My question is: do our data encourage us to notice them? 

I suspect they do not. As such, the consequence is that some of the students with the greatest capacity for educational gain receive the least pedagogical attention. Universities rightly identify students whose continuation is at risk. We devote much less effort to identifying students whose attainment could be transformed through relatively modest changes to teaching, feedback or academic relationships.

Educational gain is not distributed evenly across a student population. Some learners require substantial pastoral support simply to remain engaged, whilst others may experience significant academic development following one timely conversation, carefully designed formative feedback or an increased sense of belonging within their discipline. If learning analytics cannot help us distinguish between those different forms of educational opportunity, then we are asking increasingly sophisticated systems to answer significantly limited questions. Perhaps the challenge facing higher education is not to build better dashboards – perhaps it is to ask more educational questions of the data we already possess.

The argument I have made in this article is not that universities should invest less in learning analytics. Institutions should continue to develop their analytical capability because understanding students’ experiences is crucial if they are to be enhanced. However, my concern is that we have become so focused on prediction that we have paid insufficient attention to pedagogy. Learning analytics has become exceptionally good at telling us which students are most likely to struggle. It has been much less successful in helping us understand how teaching itself should change. If universities genuinely want to reduce awarding gaps, I contend that we should ask different questions of our data.

First, what do the data tell us about our teaching rather than our students? Much of our current use of learning analytics encourages us to examine student behaviour. Attendance, engagement and submission patterns become indicators of individual risk. Those measures have value, but they tell us little about the effectiveness of our curriculum, assessment design or feedback practices. Every dashboard should prompt course teams to ask whether the learning environment itself requires attention before concluding that the student does.

Second, where is the greatest opportunity for educational gain? Universities have become highly effective at identifying educational risk. We should become equally interested in identifying educational opportunity. Some students require substantial pastoral intervention to remain engaged. Others may transform their academic performance following timely feedback, clearer expectations or a stronger sense of belonging. Educational gain is not distributed evenly across a student population. Learning analytics should help us understand where relatively modest pedagogical changes are likely to have the greatest impact.

Third, who have we stopped noticing because they continue to pass? The silent middle is not simply a descriptive label. It represents a challenge to institutional assumptions about success (Mather, 2026a). Students who consistently achieve marks in the middle classifications rarely appear within institutional dashboards because they generate little concern. Their continued progression can disguise unrealised potential. Universities that are serious about improving outcomes should become just as interested in students who have plateaued as those whose continuation is at risk.

Fourth, what should we teach differently tomorrow because of what the data tell us today? This is the most important question of all. Learning analytics should not conclude a conversation, it should start one. If programme teams leave meetings with a list of students to contact but no discussion about teaching, assessment or curriculum, then the data has served an administrative purpose rather than an educational one. Professional judgement remains the most powerful tool available to educators. Analytics should strengthen that judgement, not replace it.

Finally, how will we know whether our teaching has changed outcomes? Universities devote considerable resources to measuring continuation, progression and degree classifications. Those metrics are important, but they should not become proxies for educational quality. We should also ask whether students have become more confident learners, whether they participate more actively, whether they experience a greater sense of belonging (on their terms) and whether teaching has enabled them to exceed the trajectories that seemed likely at the beginning of their studies. Those are educational outcomes. They deserve to be measured with the same seriousness as institutional performance indicators.

None of the questions I have posed requires higher education to adopt a system such as ALPS. Schools, sixth form colleges and universities pursue different purposes, educate different learners and operate within different frameworks. The lesson is not the system itself – it is the educational philosophy that underpins it. Freire (1970/2005) argued that education becomes transformative through dialogue, reflection and action. The same principle should apply to learning analytics. Data should not become another mechanism for classifying students or managing institutional performance – it should encourage educators to reflect critically on their practice, challenge assumptions about learning and act in ways that improve students’ educational experiences.

My own work over the past year has explored transition, belonging, good teaching and the silent middle (Mather, 2026a, 2026b). This article has helped me realise that these are not separate conversations; they point towards the same conclusion: students succeed because teaching matters. Relationships matter, recognition matters, belonging matters and curriculum matters. Learning analytics only become educationally significant when they help us improve all of those things.

David Mather is Associate Head of School (Students) and Senior Lecturer in Educational Leadership and Management at the University of Portsmouth. His research explores identity, recognition, and transition across educational and professional contexts, with a particular interest in improving educational practice and the student experience. He is the founder of We See You, an initiative supporting Armed Forces, veteran, and Blue Light communities in higher education.


by Leave a comment

A narrator Is not a witness: what AI got right — and wrong — across three Ghanaian MBA classrooms

by Carlene Kyeremeh

It was late, and I was building slides on fiscal policy for a class of twenty, most of whom had never taken economics before. I asked an AI tool for a worked example of expansionary policy — something concrete enough to anchor the mechanism. It answered in seconds: a government cuts taxes and raises infrastructure spending; roads get built; contractors hire; demand rises. Clean. Usable. American. The mechanism travelled, but the roads were interstate highways, and the government spending the money was not one my students would ever petition. The example had arrived so quickly, and so confidently, that I nearly pasted it in. For a moment, convenience almost became curriculum.

Then, in the same session, the same tool did something I could not have done as well alone. I asked it to scaffold the vocabulary — a glossary that assumed nothing and met students who did not yet have the words. It was patient in a way I am not always patient at that hour. It found analogies; it sequenced the terms so that each rested on the one before. That slide was better for the machine’s help.

Within a single sitting, the tool had lowered the barrier for my students and quietly imported the exact default my work exists to help people notice. Across three MBA courses this term, I began to see the same pattern: the tool improved many aspects of how I prepared to teach, and became least dependable where the teaching required local knowledge, current evidence or an institution named correctly. The help landed on form. The failures landed on specificity.

In my international trade finance course, I extend a running case — Adom Naturals Ltd, a Kumasi agribusiness I built so that trade finance stops being abstract and acquires a location, a product and a currency exposure. I asked the tool to situate the firm against current conditions: what the African Continental Free Trade Area (AfCFTA) changes for a business like this, and where COCOBOD now sits in the picture. It answered in assured paragraphs and told the familiar story: Ghana grows some of the world’s finest cocoa, ships much of it out with limited processing, and watches the greater share of value accrue downstream. Fluent, orderly, and a season out of date.

The confident narrator

It missed what I happened to be holding in a government source that week: in February 2026, Cabinet directed that, from the 2026/27 crop season, a minimum of 50 per cent of Ghana’s cocoa beans should be processed locally. For Adom Naturals, that reform changes the opportunity set. Cocoa liquor, butter, cake and other processed products can retain more value within Ghana and may qualify for preferential treatment in African markets where the relevant AfCFTA rules of origin and tariff requirements are satisfied. The tool had narrated the extractive arrangement in the present tense and missed the policy intended to change it.

Nothing marked the claim as stale. The tool warns that it can make mistakes, in a line printed beneath every answer, but a caveat attached equally to everything is not calibration; it is the absence of it, dressed as candour. A colleague says: I am sure of this; check me on that. The tool says it might be wrong about anything, then says everything in the same even voice. The danger was never that it made mistakes. Every source makes mistakes. The danger was that it sounded exactly as certain when it was wrong as when it was right.

The wrong institution

Financial regulation showed me the problem from another angle. Ask a general-purpose tool about capital adequacy, disclosure or market conduct and it is fluent, because the published record is thick with Basel standards, US and UK regimes, and decades of commentary. Ask it to route the same questions through Ghana’s regulatory architecture and the fluency thins.

When I asked which body supervises an insurer in Ghana, and then which oversees a securities offering, it reached both times, confidently, for the Bank of Ghana. The answer was plausible because the central bank is prominent in Ghana’s financial system. It was nevertheless wrong. Insurance supervision belongs to the National Insurance Commission under the Insurance Act, 2021; securities-market regulation belongs to the Securities and Exchange Commission under the Securities Industry Act, 2016, as amended. When I named the specific commissions, the tool corrected itself at once. The information was retrievable; it was not the default.

Three defaults, one voice

Set the three moments side by side and a more complicated pattern emerges. The fiscal-policy example exposed a geographical default; the cocoa case, a temporal one; and the regulation case, an institutional one. These were different failures, but they arrived in the same confident voice.

I cannot inspect the tool’s training archive, so I cannot attribute every error to missing data alone. A stale policy claim may reflect a knowledge cut-off or the absence of live search. A regulatory error may reflect weak retrieval, poor weighting or the greater prominence of a general institution over a specialised one. What I can observe is an asymmetry of retrieval: general and North Atlantic formulations arrived unprompted, while Ghanaian specificity had to be named, sourced and verified into view.

That asymmetry belongs in the larger conversation about AI and epistemic justice — about whose knowledge is dense enough, accessible enough and prominent enough to be retrieved fluently, and whose is thin enough to be flattened, displaced or missed. The tool did not invent the hierarchy of whose knowledge counts. It inherited a record shaped by that hierarchy and can reproduce it at scale, in fluent prose. Better models may reduce some errors, but model improvement alone cannot repair knowledge that remains absent, inaccessible or systematically underrepresented.

I develop that argument more formally elsewhere, in work currently under review. Here, I want only to report what it looks like from inside three classrooms, at the point where defaults become examples and examples become curriculum.

Verification is the work

I use these tools daily and they earn their place, so let me be honest about the difficulty. The answer is not refusal; refusing the help is not a decolonial act, only less help. The answer is the discipline I have argued for all along, now turned on the machine: no sentence enters the curriculum until it points to a source I can hold. Verification is not the friction that slows the tool down. With a tool like this, verification is the work.

I have also begun turning that work into a learning activity. I place selected AI outputs beside the relevant primary or institutional sources and ask students to identify what the model has generalised, dated or assigned to the wrong body. Verification becomes not only my quality-control procedure but part of the curriculum itself.

Perhaps that is the graduate skill this moment now asks for. Not simply how to find information, the tool is generous with information, but how to test information whose presentation gives no sign whether it has earned our trust. The scarce skill is no longer retrieval. It is discernment. Teaching has always required two kinds of expertise: explaining ideas well, and knowing where they belong. The tool is becoming remarkably good at the first; the second is still ours. It narrates beautifully — but a narrator is not a witness, and decolonising the curriculum now includes learning to interrogate the archive that speaks back.

If your tool has ever been confidently wrong about your own institution, your own regulator or your own country’s data, I would like to know what it got wrong — and whether a student would have caught it.

Dr Carlene Kyeremeh is an Associate Professor and Vice President, University Advancement, Recruitment & Research, at All Nations University, Ghana, where she teaches managerial economics and international trade and finance on the MBA programme. Her research examines decolonial curriculum reform, gender equity and academic mobility in African higher education, with the African Continental Free Trade Area as a recurring empirical anchor. She is currently researching the reintegration of diaspora-return faculty in Ghanaian universities. She writes The Decolonized Curriculum, a newsletter on curriculum decolonisation in African higher education. 

LinkedIn [https://www.linkedin.com/build-relation/newsletter-follow?entityUrn=7412983304175534080]

Author’s note: This article is adapted and substantially expanded from Issue 13 of The Decolonized Curriculum.


by Leave a comment

Come back, Robbins, all is forgiven…

by Paul Temple

Mike Shattock and I are at present working on a book on English higher education policy from 1963 to the 2024 general election (forthcoming from Bloomsbury, since you ask). Our start date was set for us by the publication in that year of the Robbins report (Committee on Higher Education, 1963), which offers a classical model of “rational” policy formulation, identifying choices based on knowledge of various kinds, supported by testable data, and located in a defined organisational and policy environment. Distinctively, the report addressed both what universities should do, and how they should do it. Perhaps most crucially for the future of British higher education, it demolished the “pool of talent” argument, which claimed that only a fixed proportion of the population would be able to benefit from a university education, and so substantial expansion would pointlessly “dilute” the quality of existing provision. It was remarked at the time that it was indeed a happy coincidence that the size of the “pool of talent” in Britain exactly matched the number of existing university places. Taking this and other research findings, the Robbins Committee proposed a scheme of expansion of higher education in Britain that received cross-party support in Parliament, and created the structure that, in large measure, survived into the twenty-first century.

Simply to outline the work of the Robbins Committee is to invite nostalgia for a bygone age of policy-making, in Britain and probably most Western countries. Rory Stewart, a former minister in the Conservative administration of Boris Johnson, serving between 2015 and 2019, describes how he came to realise that, in different ministerial roles, his task was not to devise and implement policies in order to achieve actual performance improvements in his areas of responsibility – variously, air pollution and prisons – but to present anticipated failures in these fields in the most politically-helpful way: government as performativity (Stewart, 2023). Pathways to improvement in the areas of Stewart’s responsibilities were well-understood within the relevant professional groups, but were blocked by political and organisational obstacles, not least, as Stewart recognised, by the short-term nature of ministerial appointments and changing political priorities, set against the long timescales for change in most areas of public policy. Stewart came to realise that attempts to overcome these obstacles in the time likely to be available to him would merely highlight his political naivete. As a result, policy was enacted as performativity, which, various writers suggest, has become the dominant approach to public policy making in the UK and many other countries (Kim, 2024; Voss, 2014). The contrast with the world of the Robbins Committee is of course depressingly stark.

Performativity may also give rise to overlapping policy agendas that are not necessarily incompatible but which have been developed apparently in isolation from each other. One example is what happened following the 2011 Higher Education White Paper when new policies were intended to create dramatic changes to British higher education. But the performative approach to policy (“How can we present this in a headline-grabbing way?”) produced what has been termed “layering”, when new methods are attached to existing ones, which may change the ways in which the original structures worked (Mahoney and Thelen, 2010: 16; Ward et al, 2025). Thus, the new student fee regime was intended to change institutional methods by giving students market power; however, the institutional status hierarchy was unchanged, and student market power could only be exercised, if at all, in relation to institutions towards the bottom of the status hierarchy. A complicated fee regime was thus overlaid on an institutional structure created on a different basis.

As if that was not bad enough, the implicit (though unstated) assumption was that the students whose tuition fees were to be loaned from public sources would be attending one of the then-established universities, having met relatively rigorous entry requirements, and where considerable effort would be put into ensuring student progression and completion. That is to say, the new fees-and-loans system was to be operated within a well-understood and stable institutional environment. But simultaneously with this radical policy development, the same government was encouraging the creation of profit-making institutions, applying minimal entry requirements and with no control on student numbers and with little interest in academic achievement. These privately-owned companies would then be able to maximise the fee income paid to them on behalf of their students, which they would be free to use as they wished.

Put together, these two policies have created a situation that is financially and educationally disastrous (the same situation that existed in the United States), but no policy-makers seemed willing, at the time or subsequently, to assess the joint impacts of the two policies: each made a certain sort-of performative sense, but together – well, have you heard a coherent defence of the current set-up? Do let me know if you have.

SRHE Fellow Dr Paul Temple is Honorary Associate Professor in the Centre for Higher Education Studies, UCL Institute of Education.


by Leave a comment

Capability for special, adult, and further education

by Gavin Moodie and Leesa Wheelahan

The capabilities approach focuses on human flourishing rather than augmenting productive capacities, although the latter may be necessary for the former. It is also concerned with the potential of people to achieve or become, rather than maximising happiness or the sum total of utilities that individuals have. It is concerned with the real freedoms people have rather than economic resources, as it recognises that different individuals, communities and societies need different types and different levels of resources to live dignified lives.

Students with special educational needs and disabilities

The capabilities approach provides the basis for understanding and addressing multiple and overlapping forms of disadvantage. Tikly explained that the capabilities approach ‘implies a focus on the institutional and cultural barriers that prevent inclusion of different groups’. It has implications for resourcing because it recognises that some groups will need more resources to achieve comparable capabilities as the basis for exercising choice in their lives. This is one reason why the capabilities approach was first drawn upon by the those fighting for the human rights and dignity for those with disabilities. Terzi argued that the capabilities approach ‘is innovative with respect to the centrality of human diversity in assessing inequality’. It offers another theory to guide the development of students with special educational needs and disabilities in further and higher education. This has implications for policy, funding and governance. As capabilities are embedded in their social context and manifest differently in different contexts, they require local engagement with social partners, educational institutions and a nuanced understanding of the different needs of different students.

Education for adults and part time students

The SRHE Blog has noted the under-provision of education for adults and its editor Rob Cuthbert observed that ‘Part-time students are vanishing from the system, as funding regimes repeatedly ignore their different needs and discourage institutions from providing for them.’ Capabilities theory offers a further argument for improving provision for adult and part-time students, if one is needed.

Further education

Tikly argued that the capabilities approach ‘allows for an expanded view of the purpose of [Technical and Vocational Education and Training] as supporting the development of human capabilities and functionings that individuals, communities and society at large have reason to value’. He applied the capabilities approach to understanding the broad purposes of vocational education and the role it can play in society, and also to issues of funding and governance.

We apply human capabilities to vocational education by considering what people are able to ‘be and do’ at work and through work to realise themselves and their goals. We understand capabilities to refer to the resources and arrangements of work and the broad knowledge, skills and attributes that individuals need to be productive at work, to progress in their careers, and to participate in decision-making about work, and in their life. Vocational education students need to understand how their field of practice fits within their communities and societies, and they require the capacity to be ‘citizens’ within their field, so they can help shape its future.

We argue that vocational education develops students’ capabilities in two ways. It develops students’ understanding of the world and their capacity to act in the world. And its educational processes facilitate students making decisions about their education and how they undertake it, even within strongly framed programs.

To do this vocational education should prepare students for a broad field of practice rather than a narrowly defined occupation, or worse, narrowly defined skills aligned with work-place tasks and roles. Vocational education should also contribute to and benefit from helping students develop the building blocks of these broader capabilities.

Strong institutions

The development of capabilities is underpinned by institutional frameworks and social partnerships. This points to the importance of strong public vocational education institutions in mediating the development of capabilities and in mediating links between students and other social partners. And, this requires strong public vocational education institutions to participate in discussions and debates about the nature of capabilities, and to support their development.

Sen explained that we live and operate in a world of institutions, and that ‘Our opportunities and prospects depend crucially on what institutions exist and how they function’. He explained that institutions must be evaluated by the extent to which they contribute to capabilities and to freedom. Winch said that a vocational education framework needs stability:

“… so that it can evolve along pathways familiar to participants, and so that routes and qualifications are recognized by all stakeholders … Governments must resist the temptation to change TVET structures for short-term political benefit, and should plan for robust and long-term stable structures. Last, but not least, qualifications should contain substantial theoretical content to facilitate permeability and broad occupational capabilities.”

Capabilities provide the conceptual basis of vocational education and its qualifications, but the specific focus and content of teaching and learning and curriculum requires deep understandings of the contexts for which students are being prepared, engagement with local communities of interest, and negotiation over the outcomes. For example, the capabilities that electricians need will differ from those of childcare workers.

The extent to which individuals can realise their capabilities at work is partly related to the extent to which a workplace facilitates their agency in work, the level of autonomy and support they are offered, and where they are situated in organisational structures. Hierarchical and managerial workplaces restrict agency, whereas a more expansive approach may provide opportunities for development and growth. It is particularly salient for vocational education that workers’ capabilities as workers depends crucially on the capabilities of their work unit. Work groups have been studied extensively since the Hawthorne studies of the 1920s which identified considerable variation in the effectiveness of different work groups. A work unit that supports and encourages its members to develop their knowledge and skills develops its members’ work capabilities more than others, and according to Senge, learning organisations have competitive advantages over other organisations.

Caveats in using the capabilities approach

There are important caveats in understanding the way in which the capabilities approach can be used. Tikly explained that ‘the capability approach should not be seen as providing ready-made answers to the policy issues and challenges facing the TVET sector today’. He argued that it ‘should be seen as a way of framing issues and as a starting point for evaluating policy choices’.

The second caveat is that while the capabilities approach provides a normative framework for evaluating, assessing and providing the conditions for individual well-being and social arrangements, it does not provide explanations about the causes of ‘capability deprivation’. The capabilities approach cannot be applied in the absence of theorising about social relations of power and domination.

The third caveat is that while the capabilities approach could help renovate vocational education so that it is more holistic and developmental, it cannot on its own ‘fix’ the problems of the labour market. To do this would require a focus on the labour market itself, and not just vocational education qualifications. For example, while vocational education may provide education that helps students develop capabilities, these capabilities may not be able to be realised in workplaces that resist change, are discriminatory, or provide few opportunities for discretionary learning or for the develop­ment of autonomous practice.

The fourth, crucial, caveat, is that capabilities cannot be considered in the abstract. Capabilities are not just an individual attribute: they include the resources available to a person and their personal, social and environmental circumstances that make it possible for them to realise what they reasonably value. A problem in some of the literature is that it considers capabilities in the abstract to generate idealised lists of individuals’ attributes such as general skills, employability skills and graduate attributes. However, since these are detached from the requirements of occupations and the circumstances in which they are practised they don’t advance analysis beyond other lists of general skills and attributes.

Dr Gavin Moodie is Honorary Research Fellow at the University of Oxford’s Department of Education

Leesa Wheelahan is Professor Emerita and William G. Davis Chair in Community College Leadership Emerita at the Ontario Institute for Studies in Education, University of Toronto, and Honorary Research Fellow at the University of Oxford’s Department of Education.