Can generative AI genuinely engage pupils in geography, or does it simply make lessons busier? Jack Downs shares his MSc research into how GenAI shaped engagement across a Year 9 tectonic hazards unit, and why the technology tends to amplify good teaching rather than replace it.

Jack Downs is a geography teacher and completed his MSc in Learning and Teaching at Linacre College, University of Oxford, supervised by Dr Lauren Hammond. He is interested in how emerging technologies can be embedded in disciplinary teaching without eroding the depth of pupils’ thinking.

Bringing GenAI into the classroom

When ChatGPT arrived in schools, the debate about it polarised quickly. Optimists argued that generative AI (GenAI) could personalise learning and adapt to pupils in real time (Holmes, 2019). Critics warned of the opposite, contending that leaning on the technology risks passive, superficial learning in which pupils let the machine do the thinking (Selwyn, 2019). Geography has often been an early adopter of new classroom technology, yet subject-specific guidance on how to use GenAI well remains thin (Lane, 2025). I wanted to move the question out of the abstract and into a real classroom.

My MSc study was a mixed-methods case study across nine one-hour Year 9 lessons on tectonic hazards at a secondary school in England. I measured engagement using a tailored version of Martin’s Motivation and Engagement Scale (MES) (Martin, 2007) before and after the unit, gathered pupil-voice surveys after every lesson, kept a teacher reflective log, and interviewed geography colleagues. Throughout, I treated engagement as multidimensional (Fredricks et al., 2004), spanning behavioural, emotional and cognitive dimensions, alongside agentic engagement, the influence pupils exert on their own learning (Reeve, 2012).

The intervention

I designed nine GenAI activities around a single principle. Pupils should have to explain, critique and revise the AI’s outputs, not simply receive them. Each activity used a bespoke ChatGPT “GPT” that I had prompt-engineered for the lesson, with the school’s GenAI policy and the Department for Education’s “Keeping Children Safe in Education” (2024) guidance loaded in so that every response met the relevant safeguarding protections.

The tasks offered structured choices, adaptive scaffolding and instant formative feedback. In one lesson, pupils generated an image from their own written description of a volcanic eruption, then interrogated how well it matched the geography. In another, they wrote an exam answer and received coaching and specific next steps from a feedback GPT, completing the whole feedback loop within the lesson. Elsewhere, they interviewed a GenAI “tsunami survivor” using questions they devised themselves and designed an earthquake disaster plan collaboratively in groups.

What the numbers showed

The MES compares eleven dimensions of motivation and engagement before and after the unit. Two adaptive dimensions rose clearly. ‘Self-belief’ increased by +0.47, meaning pupils were more confident they could complete their work and do well, and ‘task management’ rose by +0.38, suggesting they organised themselves more effectively. Instant, specific feedback appeared to be doing real work here.

The picture was not uniformly positive, though, and that is what makes it interesting. ‘Valuing’ (-0.19), ‘learning focus’ (-0.12) and ‘persistence’ (-0.39) all dipped. More strikingly, the three largest changes of all were maladaptive. ‘Anxiety’ and ‘uncertain control’ each rose by +0.54, and ‘self-sabotage’ by +0.56. The rise in anxiety clustered around examinations, and the rise in uncertain control around whether the feedback pupils received genuinely helped them improve. Confidence and organisation went up, but so did unease.

What pupils said

Pupils’ own responses were overwhelmingly linked to affective engagement. “Fun” appeared 94 times across the surveys and “interesting” 45 times. The exam-feedback activity was the most popular of all, with nineteen of twenty pupils saying it made the lesson more interesting. One wrote that “the next steps really showed me how to improve”; another that “I was determined to improve using what the AI told me”. Several described the tool in relational terms, with one saying, “I interact with the AI like a friend, and it gives me hints and it tells me when I’m wrong or not.” Where activities offered genuine choice, pupils valued the ownership, noting that it was “fun to interview a [tsunami] survivor and come up with my own questions”.

Engagement or novelty?

This is the honest heart of the study, because enjoyment is not the same as learning. Where a task demanded reasoning, the exam-feedback activity above all, pupils showed genuine metacognitive reflection and built a firmer understanding rather than just being handed an answer. Where a task did not, the gains were shallower. In the image-generation lesson, some pupils wrote the barest description they could and rushed to finish, which points to novelty rather than deep thinking. A number of pupils, given the chance, simply used GenAI to obtain answers, and lower-attaining pupils sometimes needed more scaffolding before they could access a task at all. Cognitive engagement was the hardest dimension to secure, and it depended almost entirely on how the task was designed.

What it means for teaching geography

The central finding is straightforward. GenAI does not generate engagement on its own. It amplifies well-designed pedagogy when its integration supports autonomy and feedback, and it can just as easily amplify passivity when it does not. Three practical principles follow from the study.

Turn “fun” into thinking. Where pupils enjoy a GenAI task, build in a requirement to explain and justify their choices. After generating an image of a volcanic eruption, for instance, ask them to account for the geographical processes involved and link them to a case study, shifting the work from description to reasoning.

Offer autonomy within a framework. Pupils appreciated choice but felt lost when options were wide open, which is part of what drove the rise in uncertain control. Suggested pathways of varying difficulty preserve a real sense of volition while giving pupils something to hold on to.

Humanise the feedback. Pupils valued the AI’s guidance but still looked to their teacher for validation. A brief teacher-led debrief after each feedback cycle, drawing out common misconceptions and linking them to disciplinary ideas, helps develop competence and relatedness that the machine alone does not provide.

Looking ahead

This was a single-site study with a small sample and a short timeframe, so the findings are indicative rather than definitive. Even so, they suggest that geography, with its long history of adopting new tools, is well placed to shape the use of GenAI thoughtfully rather than reactively. The prize on offer is not novelty but deeper, more autonomous engagement, and whether we reach it depends far less on the technology than on the pedagogy we build around it.

References

Downs, J. and Campbell, T. (2024) Using AI to transform adaptive teaching at A level. Teaching Geography, 49, 102–104.

Fredricks, J. A., Blumenfeld, P. C. and Paris, A. H. (2004) School engagement: potential of the concept, state of the evidence. Review of Educational Research, 74, 59–109.

Holmes, W., Bialik, M. and Fadel, C. (2019) Artificial Intelligence in Education: Promises and Implications for Teaching and Learning. Center for Curriculum Redesign: Boston, MA, USA. https://circls.org/primers/artificial-intelligence-in-education-promises-and-implications-for-teaching-and-learning

Lane, R. (2025) Mitigating risks, embracing potential: a framework for integrating generative artificial intelligence in geographical and environmental education. International Research in Geographical and Environmental Education, 1–18. DOI:10.1080/10382046.2025.2458561

Martin, A. J. (2007) Examining a multidimensional model of student motivation and engagement using a construct validation approach. British Journal of Educational Psychology, 77, 413–440.

Reeve, J. (2012) A self-determination theory perspective on student engagement. In: Christenson, S. L., Reschly, A. L. and Wylie, C. (eds.) Handbook of Research on Student Engagement. Boston, MA: Springer.

Ryan, R. M. and Deci, E. L. (2000) Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55, 68–78.

Selwyn, N. (2019) Should Robots Replace Teachers? AI and the Future of Education. Cambridge: Polity Press.

Professor John Morgan

On Thursday this week (30.01.25), I received two messages. The first was to announce the publication of the Handbook of Geography Education, edited by Sarah Bednarz and Jerry Mitchell. The second was a reminder that Norman Graves was celebrating his 100th birthday. It was a fitting coincidence because Norman Graves was one of the founders of the field of geography education research during his tenure at the University of London Institute of Education (now UCL). Although there are few direct references to Graves’s work (about a dozen), the Handbook reflects the work he did to define the problem-space of the field.

Elsewhere, I have often remarked that the making of modern geography education at least in the UK can be dated to the publication of three founding texts. These are Graves’s (1975) Geography education, David Hall’s (1976) Geography and the geography teacher, and Bill Marsden’s (1976) Evaluating the geography curriculum. Published around the same time, what they had in common is that they were attempting to make sense for teachers of the convergence of curriculum theory and geographical theory. Geography was going undergoing its conceptual revolution and education was being reshaped by new findings from educational psychology and the introduction of rational curriculum planning. Graves summarised the challenges for geography teachers, who must make sense of the philosophical basis of their teaching and consider psychological questions of how young people learn. He noted that there were sociological questions of why some children learned faster and more effectively, but this was not his prime concern (even though it was of growing interest to others and spawned the birth of the ‘new sociology of education’ (Young 1971) also at the Institute of Education (IoE).

Graves offered a pragmatic solution to the question of the overarching conceptual framework or ‘paradigm’ with which to work. He suggested there were three such paradigms vying for attention in geography in the 1970s. These were the spatial organisation paradigm, ecosystem paradigm, and critical social science. He stated his preference for the ecosystem paradigm as most effectively allowing teachers to make the transition from the old regional paradigm to the new geography in school. He was unclear about the value that critical social science or radical geography might have for the subject in schools.

In passing we should note that Graves advocated and set the standard for using educational research as the basis for rational curriculum planning. In this he was very much a product of the time. Graves joined the IoE in the mid-1960s the time of the expansion of teacher training and educational research. He introduced the landmark MA in Geography Education in 1978. Dennis Lawton had been appointed as the first Professor of Curriculum Studies in 1964, and Harold Rosen and James Britton led the English section. They set about modernizing the training of teachers in the face of comprehensive schooling and curriculum change. Graves’s 1975 book Geography education was based on these courses. This was the first shift towards curriculum theory in geography education and his follow up – Curriculum Planning in Geography (1979) – defined the curriculum problem for teachers and offered practical advice on how to solve it.

Graves, like all of us, was shaped by the politics of the time. He took up his role at the IoE at the height of the progressive educational settlement, a time of optimism about the prospects for improving learning for all, and teachers and teacher educators enjoyed relative autonomy.  The frontiers of knowledge about the organization of learning were expanding, and universities were trusted as the best place to codify and pass on this knowledge. Those who worked with Graves took on and developed many of the ideas (Ashley Kent, Eleanor Rawling, Mick Naish, Frances Slater and a little later, David Lambert). By the late 1970s this consensus was beginning to breakdown and education policy took on a harder edge. Graves wrote increasingly about the need to defend teacher education (this is part of a longer story for another time).

For now, I simply want to acknowledge that what we recognise as the field of geography education research owes much to Norman Graves who effectively defined the problems of the field. Things have moved on, of course, and by the 1990s teacher education was increasingly defined by technocratic approaches and the critical turns in academic geography posed new challenges for curriculum theory in geography education. The internationalization of the field, something Graves encouraged and was part of, has brought with it new ways of writing and thinking about geography education, as reflected in Bednarz and Mitchell’s impressive Handbook of Geography Education.

The publication of this major contribution to geography education research on the centenary of one of its founders reminds us that we have always been, and remain, Gravesian in some way, whether this is explicitly acknowledged or not.

References

Bednarz, S. and Mitchell, J. (2025). Handbook of geography education. Springer.  https://link.springer.com/referencework/10.1007/978-3-031-72366-7

Graves, N.J. (1975). Geography in Education. London: Heinemann  Educational Books.

Graves, N. J. (1979). Curriculum planning in geography. London: Heinemann  Educational Books.

Hall, D. (1976). Geography and the geography teacher. Unwin Education Books.

Marsden, W.E. (1976) Evaluating the geography curriculum. London: Oliver and Boyd.

Young, M. F. D. (1971) (Ed) Knowledge and Control: new directions for the sociology of education. London: Collier-Macmillan