AI in education: from human-centred to more-than-human-centred

My entire career, I regarded human-centred design as an intrinsically good value. Now it’s not so simple…

Since my masters in Ergonomics, and then as a PhD student grounded in the Human-Computer Interaction community, human-centred design was a no-brainer: it was an unquestioned assumption that this was an intrinsic good. I made it a mission to bring that orientation into my work on learning analytics and AI in education. Now, I haven’t just decided that we should exclude stakeholders from having a meaningful voice in shaping the tools they’re going to be using! This has proven in our work to be empowering, ethical and effective.*

But in recent years, I’ve been learning a lot from colleagues who work within larger frames than only the quality of the process and outcome of humans interacting with software. One of those is Sharon Stein (University of British Columbia), with whom I’ve been having fascinating conversations.

In this new paper, we engage in a ‘cartography’ of five senses in which the turn to “more-than-human” intersects with how we frame AI in education. We open with this:

“Humanity is grappling with deepening ecological, political, social and technological disruptions, which operate as intersecting causal loops and are experienced as what some term a “global polycrisis” [1]. Growing recognition that a narrow focus on human interests reflects a broader human exceptionalism that has contributed to many of these disruptions has prompted renewed interest in more-than-human-centred research and design. Whereas human-centred design asks how technology can serve human needs, values, and goals, more-than-human-centred design aspires toward approaches that are less anthropocentric and more attentive to beings, relations, and material conditions that human-centred approaches have often left in the background, as well as to habituated patterns of perception and interpretation and the consequences of computational mediation [2, 3].”

Our hope then, with a cautionary note, is to “invite further inquiry into what a more-than-human turn in AIED might interrupt, newly enable, or inadvertently reproduce.”

This will be a contribution to a new workshop that we’re trialling within the EdTech/AIED community, this month in Valencia at ECTEL 2026, to open up dialogue on expanding the dominant foci of these communities: Critical and More-than-Human Perspectives on AI in Education.

Curious to know what you think…

Sharon Stein & Simon Buckingham Shum (2026). Mapping More-than-Human Approaches to AI in Education. International Workshop on Critical and More-than-Human Perspectives on AI in Education, 21st European Conference on Technology Enhanced Learning (ECTEL’26), Sept. 14-18, 2026, Valencia, Spain. http://dx.doi.org/10.2139/ssrn.7401558

Abstract: Amid deepening systemic disruptions (ecological, political, social, and technological), “more-than-human” perspectives are challenging human-centred thinking in many fields. Recent developments in AI have made human-centred assumptions newly difficult for education to take for granted, unsettling familiar assumptions about intelligence, agency, authorship, and distinctively human capacities. We consider five ways in which this reorientation can intersect with AI in education (AIED), each of which raises distinct educational possibilities and questions for AIED research, pedagogy, and design: (i) AI for learning about more-than-human intelligence; (ii) AI inspired by more-than-human intelligence; (iii) AI as a more-than-human assemblage; (iv) AI as a more-than-human interlocutor; and (v) AI as a provocation to unsettle “the human.”  We welcome the possibilities these approaches open while cautioning that a more-than-human turn does not necessarily escape the inherited habits of separability, mastery, hierarchy, and extraction that have shaped both human-centred education and technology.  Rather than proposing a settled design direction for more-than-human-centred AIED, we offer this cartography as an invitation to collectively examine the possibilities and tensions these approaches bring into view, including what inherited assumptions and separations they may carry forward. We conclude by posing several questions to deepen and extend this dialogue, rather than try to prematurely resolve it.

* for example…

Deliberative Democracy for student/staff consultation on EdTech Ethics

Buckingham Shum, S., Ferguson, R., & Martinez-Maldonado, R. (2019). Human-Centred Learning Analytics. Journal of Learning Analytics, 6(2), 1-9. https://doi.org/10.18608/jla.2019.62.1 

Buckingham Shum, S., Martínez-Maldonado, R., Dimitriadis, Y., & Santos, P. (2024). Human-Centred Learning Analytics: 2019-24. British Journal of Educational Technology, 55(3), 755-768. https://doi.org/10.1111/bjet.13442

Alfredo, R., Milesi, M., Echeverria, V., Gašević, D., Buckingham Shum, S., Zhao, L., Yan, L., Jin, Y., Fan, J. X., Pammer-Schindler, V., Swiecki, Z., & Martinez-Maldonado, R. (2025). Co-designing AI-powered learning analytics: bringing students and teachers together. International Journal of Educational Technology in Higher Education, 22(1), 28 pages. https://doi.org/10.1186/s41239-025-00572-8 

Echeverria, V., Zhao, L., Alfredo, R., Milesi, M. E., Jin, Y., Abel, S., Fan, J. X., Yan, L., Dix, S., Wotherspoon, R., Li, X., Jaggard, H. A., Osborne, A., Buckingham Shum, S., Gasevic, D., & Martinez-Maldonado, R. (2025). TeamVision: An AI-powered Learning Analytics System for Supporting Reflection in Team-based Healthcare Simulation. Proceedings of CHI25: ACM Conference on Human Factors in Computing Systems,  Article 309, pp. 1-22. https://doi.org/10.1145/3706598.3713395

De Liddo, A., Anastasiou, L., & Buckingham Shum, S. (2026). Human/AI Collective Intelligence for Deliberative Democracy: A Human-Centred Design Approach. In E. Pournaras, S. Majumdar, C. I. Hausladen, & D. Helbing (Eds.), Handbook of Democracy in the Era of Artificial Intelligence. Edward Elgar Publishing Ltd. https://doi.org/10.48550/arXiv.2603.16260 

Alvarez, C. P., Martinez-Maldonado, R., & Buckingham Shum, S. (2020). LA-DECK: A Card-based Learning Analytics Co-design Tool. Proceedings LAK20: 10th International Conference on Learning Analytics & Knowledge,  63-72 (ACM Press). https://doi.org/10.1145/3375462.3375476

UTS:CIC (Aug. 2014 – Aug. 2026)

The University of Technology Sydney’s Connected Intelligence Centre launched 1stAugust 2014, when I moved down under, after close on 19 years at The Open University’s Knowledge Media Institute. Two years ago, I marked CIC’s 10th Birthday, but today this chapter came to a close. I’ve had the privilege of 12 years leading this amazing team in a research-active innovation centre within the VP/DVC Academic’s Education Portfolio. My thanks to Shirley Alexander for conceiving CIC, and Kylie Readman who continued to recognise our work.  I now return to my substantive position as an academic, moving out from a central unit to a new academic home. But not just any home…

Those of you who’ve been in Sydney or closer to developments at UTS over the last few years will be aware of all the organisational drama, which I won’t rehearse here. Suffice to say we are reconfiguring for the future. As such processes always are, this has been costly to us all, those who have left, and those still here.

If we glance in the rear-view mirror, I think it’s not too much to  conclude that CIC made its mark within UTS, and well beyond. The organisational dynamics of getting a hybrid R&D&Service centre like CIC to work are interesting. In some of my writing I’ve tried to capture the DNA of our human-centred approach to co-designing advanced data science and generative AI tools used by many tens of thousands of students, upskilling academics who aligned their learning design and assessments to integrate the tools into a coherent student experience. A very partial snapshot:

    • Launching the Master of Data Science & Innovation, the transdisciplinary data science program that CIC coordinated from 2015-18. The Victorian Blackfriars building that was our home for many years became a very special incubator of the next generation of data scientists now embedded in so many teams across diverse sectors.
    • Instant writing feedback since 2016 using the pre-GenAI technologies of the day, and the world’s first feedback on reflective writing.
    • Customised feedback emails to each student in sometimes large cohorts of hundreds, tailored to what they had accomplished, their goals, or their learning pathway.
    • Embodied teamwork analytics augmenting our simulation wards with sensors to pick up who’s doing what and where in a nursing team exercise, with automated visualisations ready to support the immediate debrief.
    • Multiple GenAI-powered applications for educators, students and researchers, developed in close partnership with our technology, analytics and AI colleagues.
    • All of them pedagogically-grounded tools, deployed in our degree programs, with empirical evaluations. Rigorous research published in the leading journals and conferences that put UTS on the international research and practice map in Learning Analytics and AI in Education.
    • Pioneered the use of Deliberative Democracy for student/staff consultation on AI/EdTech Ethics
    • Listening not only to educators on the challenges and opportunities of GenAI, but also to the diversity of student voices.
    • PhD alumni who’ve conducted Design-Based Research in partnership with faculties to develop co-design approaches, iterate next generation feedback tools, and advance conceptual frameworks and methods.
    • And so much more — see the website and news stories.

To all the students, academics, and professional staff who’ve been part of the CIC journey — your dedication, expertise and teamwork made the magic happen — Thank You! CIC has been a significant chapter in many lives 🙂

But now the page is turning, and we embark on the next chapter. I’m delighted to announce that I and my PhD students are crossing the few steps from the UTS Tower over Alumni Green, to join the Transdisciplinary School. TD School is where you’ll find the most extraordinarily eclectic group on campus, inhabiting the edges and liminal spaces where disciplines rub shoulders (not always comfortably), working with stakeholders on their most pressing societal challenges, breaking new ground pedagogically with students, and advancing knowledge about a world that isn’t carved into disciplinary boxes. “Transdisciplinarity is… when great minds don’t think alike”.

Some of you will have registered a shift in my work over the last few years which makes this move particularly significant. While the warnings of planetary overshoot have been there for decades, for those ready to listen (Limits to Growth anyone?), it’s becoming quite clear every passing month how serious the disruption to life as we know it is going to be. This is not just a climate crisis, but a deeply entangled set of ecological, societal, political, technological, psychological (and hence unavoidably educational) systems interactions. The global polycrisis is just one of the latest names to bring deep systems thinking and change strategies to our attention. The move to TD School opens fresh conversations and trajectories, as I figure out the work I am called to do, and who to do it with. I already work closely with some TD colleagues and PhDs (indeed two CIC alumni are now faculty: Simon Knight and Antonette Shibani), and so I very much look forward to seeing what emerges next.

And if you wish, do please share a CIC memory/reflection…

JIME is 30! Editorial retrospective

Cast your mind back to 1996… the Web’s shiny and new, and we’re trying to figure out what this means for scholarly publishing in EdTech…

The Web is a 3 year old toddler, we’ve just figured out how to embed multimedia in Mosaic and Netscape browsers (hello Apple QuickTime, MacroMind Shockwave, Java applets…), Paul Ginsparg’s 5 years into something he calls arXiv, Stevan Harnad‘s championing open access journals and piloting “scholarly skywriting”, and the threaded discussions of the wildly successful Usenet newsgroups have made it into Perl-based web interfaces like Daniel LaLiberte‘s HyperNews.

It’s 1996, and Diana Laurillard has just invited Tammy Sumner and me to serve as inaugural editors of a new journal she’s conceiving, called JIME: The Journal of Interactive Media in Education. As she now recalls in the 30th Anniversary Guest Editorial, just out:

“The interactive digital experiences we were creating for learners were so exciting, so different from the teacher-controlled sequence of explanations with multiple choice tests that abounded then, and still do. The point of using digital media was that the students could interact with a model of something, explore it, find out what it could do, use it to test their own thinking – the design was to be learner-led.

So the big idea for a new journal was to be wholly online, so that you could link in an interactive simulation or demonstration of your wonderful new interactive medium, to enable people to experience your ground-breaking creation.”

Tammy and I had just joined The Open University‘s new Knowledge Media Institute, so we saw an action research opportunity:

“What would happen if we fused all these new affordances, and stretched things further? Interactive multimedia embedded in the article, all within the web browser, plus open access publishing, plus (inspired by computer-supported discourse) review discussions, continuing with open commentaries after publication? Interactive not only with regard to the multimedia, but the scholarly discourse.”

Well, as my blog posts about JIME reflect while I was an editor, and the many articles in the decades that followed, JIME was here to stay. Many thanks to JIME’s current editors Robert Farrow & Katy Jordan for conceiving this celebration editorial by all of JIME’s editors from the last 30 years who have stewarded it through those rapids. Quite a case study in Open Access scholarly publishing!

Enjoy their reflections 😀

Laurillard, D., Buckingham Shum, S., Sumner, T., Scanlon, E., McAndrew, P., Jones, A. and Weller, M. (2026). 30th Anniversary Guest Editorial: Reflections from JIME Editors. Journal of Interactive Media in Education,2026(1): 15, pp.1-12. https://doi.org/10.5334/jime.1163

Figure 1: Interactive multimedia embedded in JIME articles rendered in a web browser. (1) Interactive extract from an art history CD-ROM that enables readers to play with the construction of a painting (Durbridge & Stratfold, 1996); (2) interactive Java applet for visualizing code execution over the internet (Domingue & Mulholland, 1997); (3) extract of video showing children programming a robot (Repenning, et al. 1998); (4) introductory walkthrough of an economics package with commentary from the author (Soper, 1997).

Figure 2: JIME’s document interface. On the left is the Article Window, on the right the Commentaries Window showing the top-level outline view of discussion about the document. Key: (1) Comment icon embedded in each section heading linking to section-specific comments; (2) active contents list extracted from the section headings; (3) print versions as HTML and PDF; (4) numeric or author/date citation automatically linked to corresponding reference in footnote window; (5) a reverse hyperlink is inserted for each citation of a reference; (6) an editorial note to draw attention to a controversial issue in the author–reviewer debate that ‘made it’ into the published version; (7) section-specific review comment; (8) an editorial comment summarizing the review  discussion and  specifying change requirements. (Note that there are two versions of the user interface: one as shown, and for smaller displays, the document and discussion are placed in separate browser windows.)

Figure 3: The JIME discourse-centric review process. This cycle includes private and public open peer-review phases, with active stakeholders involved at different points.

GenAI’s disruption of HigherEd — for students

There’s been a huge amount written about how the emergence of generative AI has disrupted the lives of school teachers and university academics. Not in any way to downplay this, we’re rectifying the rather less attention that’s been paid to the impact all this is having on students.

In 2024, a consortium of Australian universities set out to understand their experience — or rather, experiences — since what has become very clear is that there is no single ‘student voice’ when it comes to attitudes to, and behaviours around, GenAI in all its diverse forms. Sponsored by the DVCs Education/Academic/Students at UTS, UQ, Deakin and Monash Universities, the UTS team is led by CIC Director Simon Buckingham Shum, with Lisa Lim, Antonette Shibani, Mohsen Ebrahimzadeh and Jan McLean.

Check out the Student Voices on AI in Higher Education project to learn more about we’ve been hearing from >8000 students who took our survey and/or engaged in focus groups: HEDx conference panels/executive briefings, presentations, stories (The Conversation, Future Campus), open source surveys, and research papers.

A fresh snapshot with an even larger sample has just been gathered to enable us to compare what’s changed in the last two years, and to investigate emerging topics. Watch this space as we crunch the data…

The headlines:

1. Students are using AI in varied and selective ways

2. Students actively negotiate integrity and ethical responsibility

3. AI is also an emotional and institutional experience

4. Students value AI feedback, but teacher feedback remains distinctive

5. AI use cannot be separated from pedagogy

6. Student partnership can shape more usable institutional responses

Selected research papers underpinning these…

Fawns, T., Bearman, M., Corbin, T., Henderson, M., McLean, J., Matthews, K. E., Oberg, G., Liang, Y., & Walton, J. (2026). Illuminating complex student realities of artificial intelligence through an entangled pedagogy framework. Higher Education (Published online: 24 July 2026). https://doi.org/10.1007/s10734-026-01730-1

Oberg, G., Liang, Y., Bearman, M., Fawns, T., Henderson, M., & Matthews, K.E. (2026). Feeling AI: Circulating emotions, institutional climates, and moral boundaries in student use of AI. Higher Education, 1-19. (Published online: 28 March 2026) https://doi.org/10.1007/s10734-026-01658-6

Chung, J., Henderson, M., Slade, C., Liang, Y., Pepperell, N., Corbin, T., Walton, J., Yu, A.S., Bearman, M., Buckingham Shum, S., Fawns, T., McCluskey, T., McLean, J., Oberg, G., Seligmann, A., Shibani, A., Bakharia, A., Lim, L.A., & Matthews, K. E. (2026). The use and usefulness of GenAI in higher education: Student experience and perspectives. Computers and Education Open, Volume 10, June 2026, 100347. https://doi.org/10.1016/j.caeo.2026.100347

Bearman, M., Fawns, T., Corbin, T., Henderson, M., Liang, Y., Oberg, G., Walton, J., & Matthews, K. E. (2025). Time, emotions and moral judgements: how university students position GenAI within their study. Higher Education Research & Development, 1–15. (Published online: 11 Nov 2025) https://doi.org/10.1080/07294360.2025.2580616.

Henderson, M., Bearman, M., Chung, J., Fawns, T., Buckingham Shum, S., Matthews, K. E., & de Mello Heredia, J. (2025). Comparing Generative AI and teacher feedback: student perceptions of usefulness and trustworthiness. Assessment & Evaluation in Higher Education, 51(5), 863–878. https://doi.org/10.1080/02602938.2025.2502582.

Meg Price (PSN 2026) Conversational AI and the Quiet Erosion of Relational Capacity

You’ve already had a conversation with AI that you used to have with a human. What did that cost you?

This was the focus of a talk by PhD researcher Meg Price (TD School) who just presented a snapshot of her work at the Annual Conference of the Possibility Studies Network. Her very transdisciplinary work  is co-supervised by academics from different backgrounds to my own, Barbara Doran and Paulina Larocca.

The abstract of her talk is below, and check out this interactive website version of her presentation. Dig further into her work on her Substack research blog.

What Are We Losing When It Gets This Easy?  Conversational AI and the Quiet Erosion of Relational Capacity

You’ve already had a conversation with AI that you used to have with a human. What did that cost you?

Millions of people now turn to generative AI, ChatGPT, Claude, Copilot, not just for task completion, but for emotional processing, relational advice, personal sense-making, and everyday decisions that were once distributed across human relationships. This shift is largely happening beneath conscious awareness: not in dramatic moments, but quietly, one prompt at a time.

This paper draws on preliminary research suggesting that these habits may narrow the field of relational possibilities. Meaningful human connection depends on conditions that conversational AI structurally removes: otherness, mutual risk, embodied co-presence, and the friction of disagreement, withdrawal, and repair, as well as the serendipity of encounters we could not have planned.  These are not obstacles to a relationship; they are the medium through which relational capacity develops. When the default response to uncertainty becomes “ask AI,” we may be practicing something quite different: a relational style that is smooth, low-effort, and low-stakes. What does this mean for our capacity to be creative, to flourish as the deeply relational, embodied, complexity-tolerating beings that we are? What possibilities close, quietly, as these habits consolidate?

These questions demand urgent, real-world investigation before patterns become too entrenched to examine clearly. To investigate, a Living Lab methodology has been developed precisely for this. Combining body mapping, guided AI reflection conversations, physiological monitoring, and facilitated group dialogue, it treats the body-mind as primary data and attends to relational capacity as it shifts in the texture of everyday life, not laboratory conditions. We draw on preliminary findings from our initial Living Lab sessions and expert interviews with researchers in relational neuroscience, attachment theory, and AI design to offer early and timely observations about what this methodology reveals and what it suggests about the stakes involved.

Coauthorship Integrity: Reconceptualising Assessment Validity for the Age of GenAI

Don’t blame the AI: you’re accountable for your work! VivaBuddy is designed to help students check they haven’t outsourced too much thinking when they co-author with AI.

Mohsen Ebrahimzadeh is a doctoral researcher in the Transdisciplinary School, supervised by Antonette Shibani and CIC’s Simon Buckingham Shum, supported by CIC Specialist Software Engineer Andrey Inkin.

We have just published a paper in a journal special issue devoted to the challenge of Generative AI and new systems of learning for Higher Education:

Ebrahimzadeh, M., Shibani, A., & Buckingham Shum, S. (2026). Coauthorship Integrity: Reconceptualising Assessment Validity for the Age of Generative Artificial Intelligence. Computers and Education: Artificial Intelligence, 10, 100609. https://doi.org/10.1016/j.caeai.2026.100609

As we move into an era when undetectable AI is both deepening and undermining thinking and writing, regardless of how much AI may have been used, Mohsen’s PhD asks students: Are you sure you understand your own writing?

This begs the question: How could we assess this, especially at scale with larger cohorts, and about any topic?While we certainly encourage teaching teams to meet with students to verify their progress and authorship, there are clearly limits to what is humanly possible. As students juggle their studies with work and carer responsibilities, across timezones, flexible 24/7 interactive orals are clearly not possible. It’s here that GenAI opens new possibilities.

To explore these, the team is developing a prototype codenamed VivaBuddy, a web app providing a formative assessment and feedback experience (a hybrid, part viva voce and part text comprehension test). It’s powered by UTS’s secure GPT hosted in Microsoft Azure, with a custom user interface developed in CIC. Mohsen’s teaching and assessment design expertise, combined with his newly acquired prompt engineering skills, have enabled him to develop a prototype that generates a range of tailored questions to verify comprehension of any text it is given. These questions are in different formats and of varying difficulty, according to Bloom’s revised taxonomy. Students receive feedback at the end, and are then offered the opportunity to open a conversation if they want.

Within the evidence landscape of assessment validity, Coauthorship Integrity is a proposed new category of validity evidence, with AI vivas such as this a new way to gather evidence of learning, firstly in formative scenarios on draft writing, but potentially leading to summative scenarios.

Pedagogical walkthroughs and interviews with educators and assessment experts suggest this model holds promise for both formative and summative use.

Screenshots of the VivaBuddy prototype illustrate the custom user interface generated dynamically by the LLM, about to be evaluated by students:

 

Congrats Tommaso Armstrong, PhD!

Tommaso Armstrong has just gained 3 new letters after his name, with a human-centred design thesis on how queer people experience social and dating platforms designed largely by and for straight people…

Tomm Armstrong has a long history with CIC, starting out as an intern while studying UTS’s groundbreaking double-degree  in the Transdisciplinary School — the Bachelor of Creative Intelligence & Innovation. He developed a learning analytics dashboard for our blogging platform for the Master of Data Science & Innovation students, when CIC was running this 2015-18. And then he returned for a PhD…

Tomm’s PhD was based in the Interaction Design Discipline in the UTS Faculty of Engineering & IT, lead supervisor Elise van den Hoven. CIC’s Simon Buckingham Shum was delighted to be invited to join the team midway.

So what was his PhD about? It is now well established that LGBTQ+ (“queer”) young people depend heavily on mainstream social media and dating platforms to find peer support and explore their identities. However, we have a poor understanding of their experiences of such platforms, which are designed largely by and for straight people. Tomm’s research decided to focus on the experiences of queer young men, in order to empower them to envision future designs that recognise the complexity of their identities, their networks, and personal safety. His website not only displays his visual design skills, but conveys his PhD in an engaging way, including a site tuned for UX designers.

A paper from his research in the premier Designing Interactive Systems conference conveys the quality of the work:

Armstrong, T., Leong, T. W., Buckingham Shum, S., & Van Den Hoven, E. (2024). “This is the kind of experience I want to have”: Supporting the experiences of queer young men on social platforms through design. Proceedings of the ACM Designing Interactive Systems Conference, Copenhagen DK. ACM, New York, NY, USA, pp. 1681-1700. https://dx.doi.org/10.1145/3643834.3661564

His full thesis is in the UTS OPUS archive (see below). Tomm is now based in San Francisco — connect with him to at https://tomma.so

Designing to improve social platform experiences for and with queer young men

Tommaso Armstrong (2025), Doctoral Dissertation, University of Technology Sydney, AUS: http://hdl.handle.net/10453/193512

Abstract: Queer young men depend heavily on social platforms. However, their needs are often not adequately considered in the design of general platforms and they can be exposed to intra-community harms on platforms such as dating apps. Recent work within HCI has exposed some of these issues, although there are limitations with existing work understanding current experiences and there is a lack of work that involves design approaches. To address these gaps and extend current understandings, this research took a multi-study approach. First, an exploratory study used semi-structured interviews with 9 participants to ground the work. Second, an in-depth study with 24 participants used semi-structured interviews and probes to better understand experiences. Finally, a co-design study was used to translate findings into design concepts. This started by running co-design workshops with 13 technology designers before the resulting concepts were evaluated in sessions with 15 queer young men. The findings from this work extend current understandings of how queer young men use social platforms by: confirming the importance of social platforms for queer young men, showing that curation practices extend beyond concealment of identity, calling attention to intra-community harms as a significant issue, highlighting the role that dating apps played in finding community and peers, and bringing light to ways that social platforms are used in not safe for work ways not currently described in HCI literature. This work also provides a number of design recommendations in the following areas: giving people have more agency over their experiences, helping people navigate mismatched expectations on dating apps, and helping people connect to community. This thesis makes four main contributions. Empirical contributions relate to extending understandings of, the experiences of queer young men on social platforms, and how to design social platforms to be supportive of their experiences. Two further artefact contributions are made, a probe kit that can be used in future work, and a design resource for social platform designers that makes our design recommendation accessible outside of academia. The thesis ends by proposing a number of directions for future work to design platforms in ways that are: more dynamic, reduce the negative impacts of idealised presentations and support trust on dating apps; and to conduct further research into the ways that social platforms are used: to curate presentation beyond concealment, in ways that include dating apps and NSFW uses; and to explore the transferability of our findings to other groups.

Polycrisis, education, conversational AI?…

Could conversational AI serve as “thinking partners” to stretch our reasoning, deepen reflection, and foster the intellectual agility needed to navigate these turbulent times?

I’ve been giving a series of talks this year, including to the Societal Impact of AI Symposium at UNSW Sydney, and the international Online Teacher Education Conference. I was honoured most recently to contribute to The Hong Kong Polytechnic Educational Design Centre’s International Dialogues on Educational Transformation — Replay • Slides

Session Description: As overlapping global crises reshape our world, education faces a profound challenge – how do we prepare students to navigate complexity, uncertainty, and systemic collapse? In this webinar, Prof. Simon Buckingham Shum argues that conversational AI offers more than efficiency and that it can serve as a “thinking partner” to stretch our reasoning, deepen reflection, and foster the intellectual agility needed to navigate turbulent times. Drawing on his recent work, Prof. Buckingham Shum will explore how dialogical AI tools might help educators and learners engage with the deep uncertainties of the polycrisis era. Join us and don’t miss this opportunity to rethink what it means to think with AI in challenging times.”

How do successful researchers learn to become better researchers?

Yuveena Gopalan‘s doctoral research is studying how successful researchers learnt their craft. The first major publication from her PhD has just come out, distilling the insights from interviews into a conceptual framework.

Strange as it may seem, relatively little is known about how academic researchers learn. Yes there are training courses to build specific skills, but all the evidence from workplace learning studies shows that this is not how professionals in other sectors learn. Co-supervised with David Boud, Yuveena Gopalan’s thesis is investigating how successful researchers, spanning different career stages, reflect on how they got better at research, and navigated the challenges of the academic research journey.

So I’m delighted to share this open access publication distilling the results of many hours of interviews with highly published  researchers at early, mid- and senior career stages.

Gopalan, Y., Buckingham Shum, S., & Boud, D. (2025). The professional learning of academic researchers through their career. Studies in Higher Education, Published online: 19 May 2025. https://doi.org/10.1080/03075079.2025.2505932

Abstract: Professional development is necessary to sustain continual learning in any workforce, including academic researchers. However, researcher development strategies and support have been largely informed through institutional strategies, often conceived and deployed without the active participation of researchers. Several studies recognise the limitations of this approach and argue for the importance of understanding researchers’ perspectives on their learning. With an international focus, this paper examines ways in which leading researchers develop in becoming better researchers. Its distinctive contribution is to provide evidence of how academic researchers talk about their own learning, how it is conducted and what they have found effective in their careers. The paper reports the findings of a study that involved interviewing leading international researchers at three different career stages (early, mid and senior) in two fields. Four main themes were identified from the research: establishing expertise, pursuing passion, coping with challenge and change, and building belonging, with an overarching interrelationship between social and personal dimensions to learning. The findings are in line with workplace learning theories, and evidence: academic researchers, like other professions, learn predominantly through informal, unstructured and social means and are contingent on practice needs. While this alignment with our current understanding of professional learning might seem unremarkable, it has practical implications for supporting researcher development. Evidence-based approaches to examining researchers’ continued professional learning and development could promote researcher engagement and support institutional efforts to promote learning at both personal and community levels.

Exploring the Potential of LLMs for Inductive & Deductive Coding

The CIC team recently gave an overview of our recent LLM work to our Qualitative Data Analysis colleagues, as part of the Aspire QDA series

Simon Buckingham Shum (Connected Intelligence Centre), Antonette Shibani (TD School), Lisa-Angelique Lim (Connected Intelligence Centre) & Ram Ramanathan (Connected Intelligence Centre). This is work from collaborations with Aneesha Bakharia, Trish McCluskey & Nazanin Reza zadeh mottaghi

[Slides PDF]

ABSTRACT: Until recently, qualitative data analysis (QDA), such as the deductive and inductive coding of textual data, was considered the preserve of human researchers. The nuanced judgements required to apply a complex coding scheme, or to discern themes that evolve into a coding scheme, were beyond algorithms. However, the emergence and mainstream availability of large language models (LLMs: e.g., GPT, Gemini, Claude, Llama) has catalysed rigorous research into their ability to perform such QDA in minutes. This is accompanied by healthy debate on whether this could lead to the full automation of certain kinds of analysis, or the augmentation of their work through productive, hybrid analysis with a new generation of interactive QDA tools. Using LLMs hosted by privacy-respecting, secure, university instances, we have been testing LLMs for both inductive and deductive coding, and welcome your thoughts on how we address important considerations including:

  • How can we translate a theory-grounded codebook into a system prompt guiding the LLM?
  • How do we evaluate the quality of the coding compared to human researchers?
  • Since (like humans) LLMs are intrinsically variable in their coding, how do we understand and manage this variability?
  • How can an LLM provide a transparent account of its inductive coding of a corpus so humans can understand it?
  • How will human and machine analysts work together in the future, harnessing their respective strengths?
  • What concerns do researchers have about automated coding, and can these be addressed?

Publications for the details…

Bakharia, A., Shibani, A., Lim, L.-A., McCluskey, T., & Buckingham Shum, S. (2025). From Transcripts to Themes: A Trustworthy Workflow for Qualitative Analysis Using Large Language Models. Proceedings of Workshop From Data to Discovery: LLMs for Qualitative Analysis in Education, LAK25: 15th International Conference on Learning Analytics & Knowledge, Dublin, IRE, pp. 1-10. https://ceur-ws.org/Vol-3995/LLMQUAL_paper1.pdf

Ramanathan, S., Lim, L.-A., Mottaghi, Nazanin R., & Buckingham Shum, S. (2025). When the Prompt Becomes the Codebook: Grounded Prompt Engineering (GROPROE) and its Application to Belonging Analytics. Proceedings LAK25: 15th International Conference on Learning Analytics & Knowledge, Dublin, IRE. https://doi.org/10.1145/3706468.3706564

Why we need generative friction in student-AI interaction

As the evidence from generative AI in higher education starts to roll in, it is becoming clear that “frictionless” GenAI often undermines learning.

GenAI products designed to minimise the user’s effort to obtain a polished answer or document undermine intellectual work such as questioning assumptions, reframing problems, or revising claims. As this interaction paradigm establishes itself in higher education (surveys consistently report student usage at approximately 80%) the consequences of unscaffolded use of frictionless tools are becoming profoundly damaging to student learning. These consequences include weaker critical thinking, reduced metacognitive awareness, premature cognitive closure, avoidance of productive difficulty, and a shift from active participation in knowledge production to passive consumption of convergent outputs.

Baki Kocaballi (UTS Faculty of Engineering & IT) supervised a Masters interaction design project by Joseph Kizana to investigate ways of inserting interaction friction in the user interface of an ideation web app, with a preliminary evaluation by both students and professionals. This uncovered some really interesting reactions. They then  teamed up with CIC’s Simon Buckingham Shum, and Sharon Stein (University of British Columbia) to explore this further, resulting in this new paper which will be presented next month at the Workshop on Tools for Thought, as part of the prestigious ACM Conference on Human Factors in Computing Systems (CHI26).

Here’s a sneak preview of the preprint:

Kocaballi, A. B., Kizana, J., Buckingham Shum, S., & Stein, S. (2026). Drag or Traction: Understanding How Designers Appropriate Friction in AI Ideation Outputs Workshop on Tools for Thought, ACM CHI Conf., Barcelona. https://doi.org/10.48550/arXiv.2603.27550

Add your thoughts to this LinkedIn discussion…

 

AI for learner flourishing

Each year the UNESCO Mahatma Ghandi Institute of Education for Peace & Sustainable Development publishes an issue of The Blue Dot magazine, dedicated to “the relationship between education, peace and sustainable development and education for global citizenship”.

The theme for the new issue, is AI for learner flourishing. I was honoured to be invited to contribute, with a short piece sketching my current thoughts on AIED in these turbulent times, when it feels like both human and natural systems are unravelling: AI for Learner Flourishing in the Age of the Polycrisis — on the Edge of the Metacrisis (HTML / PDF of whole issue)