CSEDU 2026: Why We Should Emphasize Collaboration Over Personalization and Processes Over Products — Particularly in Online Learning

CSEDU 2026: Why We Should Emphasize Collaboration Over Personalization and Processes Over Products — Particularly in Online Learning 

The 18th International Conference on Computer Supported Education (CSEDU 2026) is part of an established annual series dedicated to advancing research and practice in technology-enhanced learning. Last month in Benidorm, Spain, the conference brought together researchers, educators, and practitioners from academia and industry to exchange insights on the evolving role of digital technologies in education. 

CSEDU serves as a global forum for discussing innovations in teaching and learning supported by computing technologies. Over the years, it has gained recognition as a platform for interdisciplinary dialogue between computer science, pedagogy, and educational policy. 

I was invited to deliver a keynote at CSEDU this year. At first glance, the term computer-supported education may sound somewhat dated — it is hard to find any form of education today that is not supported by computers in one way or another. On the other hand, one could argue that this makes the conference even more relevant. In the following, I would like to briefly summarise my keynote for those of you who were not able to attend. 

From a professor of Educational Technologies and Social Learning, you might expect praise for all kinds of educational technologies — today, inevitably, including AI — and for all the new and previously unimaginable possibilities they seem to offer. 

I will probably disappoint you on that point. Instead, I would like to focus on social learning and explain why I believe it is more important than ever, particularly in online education. 

Let me begin with my four main claims, which I will then unpack one by one. 

1. Online education is becoming increasingly important, whether in hybrid, online-first, or fully online formats. 
2. AI is neither the end of education nor a replacement for it. It is, above all, a tool. 
3. We need to prioritise collaborative learning formats over personalisation, particularly in online education. 
4. We need to place greater emphasis on process rather than product in education, especially when it comes to assessment. 

Online Education: No Longer the Exception 

In the early 2020s, during the Covid pandemic, almost everyone across Europe was forced to move to some form of online education within a matter of days. Lockdowns left very little time to develop sustainable solutions. At the time, I was working for openHPI, Europe's first MOOC platform. In addition to openHPI, our team also managed platforms for openSAP, AI-Campus, egov-Campus, and — most notably in this context — OpenWHO. 

Across these platforms, we supported around 15 million course enrolments, with OpenWHO, openSAP, and openHPI being the largest among them. 

OpenWHO, the World Health Organization's learning platform, became particularly important during those weeks. Enrolment numbers increased dramatically — from around 30,000 learners in February 2020 to 4 million by June, and eventually to approximately 7 million over the following years. At least in this respect, the WHO was well prepared for the crisis, as our collaboration had already begun in 2016. 

During the first months of 2020, I spoke with many people who urgently needed solutions and were completely unprepared. Setting up a scalable and sustainable e-learning environment takes years, not days — especially when there is a long list of requirements that cannot simply be taken off the shelf. 

As a result, many tools were introduced hastily and without any clear educational strategy behind them. When the pandemic ended, many academics and decision-makers reacted with a kind of backlash, quickly switching off most forms of digital support in order to return to an "everything back to normal" scenario. Unsurprisingly, some of these same people now claim that online education does not work. 

But we have seen — before, during, and after the pandemic — that online education can work extremely well when done properly. It can also support more formal educational formats quite effectively, not only MOOCs. 

If we take a realistic look at the state of the world — with ongoing conflicts in many regions, the possibility of another pandemic, and restricted access to education in many parts of the globe — the return to a pre-Covid, predominantly on-site model may have been overly optimistic and short-sighted. Many of the same people may soon find themselves asking for immediate help again, rather than preparing properly now. 

Online learning alone cannot solve all of these challenges. But it is a necessary component. 

Let us also consider a more positive angle. Online education enables people to study alongside their jobs, while raising children, or while caring for elderly family members. In a world where we are discussing longer life expectancy, extended working lives, and constantly changing technologies, we need to shift our focus away from the traditional 18-year-old school leaver and pay greater attention to a more experienced and settled group of students. Even if rent in university towns were affordable and commuting were feasible, an online-first or fully online university is no longer merely the second-best option — it can quite easily outperform traditional on-site universities. 

From my perspective as a teacher, this approach gives me the privilege of working with an incredibly diverse group of students. Their geographical spread can create some challenges, but it also brings remarkable cultural diversity. Many of my students have strong professional backgrounds and a wide range of life experience. The one thing they all seem to share is an exceptionally high level of motivation. 

Teaching them often feels less like teaching in the traditional sense and more like learning together — which, frankly, is the approach I prefer anyway. 

Online education is now closer to the norm than to the exception, and this trend will continue. 

Generative AI and Online Education: Powerful Tool, Overhyped Phenomenon 

Generative AI offers many opportunities, but also many challenges. In my view, it is currently heavily overhyped — particularly by those who do not fully understand it. I am not an AI expert myself, but I have seen enough hypes in education to recognise one when I see it. None of the supposedly disruptive technologies that were expected to completely revolutionise — or destroy — education have come close to doing so. My prediction is that Gen AI will follow a similar path. 

When Gutenberg invented the printed book more than 500 years ago, teachers feared for their future because they were no longer the sole source of knowledge. Distance-teaching correspondence universities in the early twentieth century did not replace universities, and neither did MOOCs in the early twenty-first century. 

According to the marketing departments of tech companies, their tools and large language models have already solved every problem we face. But if you dig a little deeper, we are still far from that. People are increasingly tired of consuming bloated Gen-AI-generated content. "If you didn't care to write it, why should I care to read it?" is a general sentiment that I’m hearing increasingly often. 

To be clear: I am not saying that AI will disappear, nor that it does not work. I am saying it is somewhat overhyped and is currently being used for many things for which it probably should not be used — just like many technologies before it. My prediction is that we are approaching the Peak of Inflated Expectations and heading toward the Trough of Disillusionment in the Gartner Hype Cycle. That doesn’t have anything to do with the actual abilities of the tools, more with the perception of the results. The good news is that beyond the hype, there are manifold use cases where AI actually helps us in our daily life to do things faster and even better. We can use AI as the powerful tool or assistant that it is, and benefit from its continuous improvement, like I did to improve the readability of this post. 

The one area where we see a disruptive influence of Gen AI are common approaches towards assessment. But to be honest, this affects only those really hard who put their bets on an approach that always was flawed. All of those remedies to fix these issues that we’re currently being “sold” on social media, are nothing new. You can find them in the theories of researchers, educators, and epistemologists such as John Dewey or Lev Vygotsky in the early 20th century, Jean Piaget and Seymour Papert in the 1950s/60s, or Jean Lavé and Etienne Wenger in the 1980s/90s.  

Collaboration Over Personalisation 

I am convinced that learning is a social endeavour and that we learn best in groups or teams — we always have. In online learning contexts, the lack of connectedness and the isolation of learners are often cited as major disadvantages. For some time now, I have therefore been using teamwork in many of my courses. 

In my research on teamwork and project-based learning in MOOCs, I found that team tasks do not necessarily improve completion rates — a topic I will not explore in detail here, except to note that the academic discussion around MOOC completion rates has always been deeply flawed. Learners who are already struggling with a course will typically struggle even more with team tasks, and many will not register for them at all, precisely because doing so would take them out of their comfort zone. 

Learner isolation makes team building more difficult — but that is also exactly what teamwork in education is meant to address. My fellow keynote speaker at CSEDU, Lluís Vicent Safont, made a compelling point in his keynote: AI will change universities, but it will not be able to replace the campus experience — the connections, friendships, networks, and informal encounters that come with it. This is precisely what we need to bring into online education. Team-based project work provides a very strong foundation for doing so. 

Lev Vygotsky and Sugata Mitra have both demonstrated, theoretically and practically, that learning alone is rarely as deep, effective, or robust as learning with some form of guidance. Vygotsky argued in the 1930s that the so-called Zone of Proximal Development can be significantly expanded when learners receive appropriate support [1]. Mitra later showed that this support does not necessarily have to come from a teacher — it can come from peers. His well-known Hole-in-the-Wall project gave children in disadvantaged communities access to computers and the internet, with older children eventually serving as tutors. His later SOLE approach — Self-Organised Learning Environments — relies on children learning together in groups with little or no direct instruction [2]. 

The idea that everyone should work at their own pace sounds reasonable at first. However, personalised learning — like gamification and other fashionable educational methodologies — can be traced back, at least in part, to behaviourism and its main proponents: Edward Thorndike, John B. Watson, and B. F. Skinner. This is a learning theory that cognitivists and constructivists have long shown to be inadequate for higher-level learning. 

In my view, a much better response than "You're faster than the others — move on to the next lesson" is: "Great, you're already done — now go and help those who haven't understood it yet." This, in essence, is the foundation of my teaching philosophy: collaboration over competition. 

John Hattie, professor at the University of Melbourne and former advisor to New Zealand's Ministry of Education, put it plainly: "The biggest problem with individualised and personalised learning lies in the overemphasis on working alone. [3]" 

Audrey Watters, a respected American technology writer, goes further, arguing that educational technologies have undermined a democratic vision of schooling because their core ideology is individualism. By encouraging students to work alone and at their own pace, they erode our shared respect for one another. Looking at what is happening in the world today, she may not be entirely wrong [4]. 

While I do not fully share Watters' tech-pessimism, I do agree with both her and Hattie's critique of personalised learning. 

This is one of the reasons why we preferred instructor-paced courses over self-paced ones in the openHPI MOOCs, and why we continue this approach in most German UDS learning programmes. It is also why I place such strong emphasis on teamwork in my classes. 

I would also like to quote David Joyner, Associate Dean of Off-Campus & Special Initiatives and Executive Director of Online Education at Georgia Tech's College of Computing — and a speaker at German UDS's 4th Conference on Digital Science in Education: 

"They join a community of classmates with a common goal [...] that drives and unites them and gives them a basis for conversation: posting on the forum, working on team projects isn't just an optional networking or socialising activity. It's a required or semi-required part of the class that drives participation. [...] The value is about the relationships they build and the networks they form, and the connections they make. [5]" 

This captures the point perfectly and aligns with my research on MOOCs that it is by far not enough to provide students with a set of communication tools and hope that they’ll start some sort of communication. It is a lot more important to give them a reason to get in touch and a topic to talk about.Process Over Product 

I mentioned earlier that the rise of generative AI has had one major impact on education: it forces us to rethink the role and form of assessment. Essays, project outcomes, papers, applications, and almost any digital artefact can no longer be taken as reliable proof of learning — if they ever truly could. Multiple-choice tests have always attracted scepticism, but today it is easy enough to build bots that handle the entire process automatically. 

My answer to this is to place greater emphasis on the learning process than on the final product. In many ways, this is what I was already doing before generative AI; I have simply had to shift the focus further. It requires additional time and effort — but this is precisely where we, as teachers, cannot be replaced by AI. Well, technically we can — but the result would be fundamentally different, and the student experience would fall far short of what it can be today. 

Scope and Themes of the Conference 

The conference focuses on computer-supported education, covering how digital tools, platforms, and methodologies shape modern learning environments. Its scope includes new educational tools, innovative pedagogical strategies, and case studies from both academic and corporate contexts. 

CSEDU 2026 was structured around several core themes: Artificial Intelligence in Education; Information Technologies Supporting Learning; Learning and Teaching Methodologies and Assessment; and Social Context and Learning Environments. These areas reflect broader trends such as the integration of AI in teaching, the rise of data-driven learning analytics, and the continued evolution of online and hybrid education models. 

Relevance in the Broader Context 

Conferences like CSEDU play an important role in shaping the field of computer science education by providing a venue for sharing research, discussing teaching practices, and building professional communities. 

At a time when digital skills and technology-enhanced learning are becoming central to education systems worldwide, events such as CSEDU highlight both the opportunities and challenges of integrating new technologies into teaching and learning. CSEDU 2026 offered a comprehensive overview of current developments and future directions in computer-supported education, bringing together diverse perspectives from across the global education and technology landscape 

[1] L. S. Vygotsky. Mind in society: The development of higher psychological processes. Harvard University Press Cambridge, Mass., 1978. (Original manuscripts [ca. 1930-1934]).  

[2] Sugata Mitra and Emma Crawley. Effectiveness of Self-Organised Learning by Children: Gateshead Experiments. Journal of Education and Human Development, 3(3):79–88, September 2014. 

[3] Hattie, John, John Hattie warnt vor falsch verstandener Individualisierung des Lernens, Online [Accessed: 24.06.2026], https://deutsches-schulportal.de/expertenstimmen/john-hattie-warnt-vor-falsch-verstandener-individualisierung-des-lernens/  

[4] Audrey Watters, Online [Accessed: 24.06.2026], https://2ndbreakfast.audreywatters.com/spot-the-difference/ 

[5] David Joyner, Online [Accessed: 24.06.2026], https://www.classcentral.com/report/educations-hidden-value