Years ago, maybe a little over twenty-two years ago, I was working as a university researcher in the field of interactive and multimedia systems. In particular, I worked for seven years on a project where we explored how to use technology in creative and constructive ways in K–12 education. It was what I called my Black Forest cake in my career; delicious and joyful.
I have been reflecting on the work we did, and adults are once again being asked to guide young people through a technology they are still learning to understand themselves. Looking back over the years I spent researching how internet technologies could be used constructively in schools, I have been struck by how many of the challenges we encountered then are reappearing today in a different form. The technologies have changed. Human nature has not.
During those years, we worked together with students and teachers, finding ways to enhance the students’ learning with the help of internet technologies and multimedia systems. I don’t really want to write about what we did, but rather about a few of the truths we discovered concerning who benefited from participating in our projects, and whether these findings might also be pertinent to the current situation of AI in schools.
I’d like to address three main learnings.
The first is that the students who benefited, meaning those who showed an increase in their engagement and confidence, were not necessarily those in the top percentile. Those students performed well and showed only a marginal increase in engagement compared with when they were having standard classes. Neither did the students in the bottom percentile show a significant increase in terms of performance. In both cases, the results were probably due to the novelty of using technology in creative ways to learn topics in their curriculum.
No, the students who benefited the most were those hovering slightly above and below the average. We researched why this was and came up with various theories. The one that struck me most had to do with the allocation of responsibility and the transparency of outputs.
When we started a project, the class would be separated into various teams. In each team, there were those who had to document the progress of the project, those who helped create the end results, and one or two who were responsible for the technology. The latter were also responsible for teaching the other students how to use the hardware and software properly.
I had a few university IT students who came with me into the classes, and we taught the “techies” and gave them clear guidelines about how to care for the devices. Then we saw to it that they were able to help their team members use the devices or software properly.
We would nearly always choose those students as our techies who were underperforming, had high absenteeism, or were somewhat introverted. Not the troublemakers, but those who were at risk of falling behind. We did not elevate them to the top of the castle; rather, we made them realise that the other students, and we in the research party, needed their support. They were the ones the other students called if they were stuck with a problem. They were also responsible for doing backups of data and collecting all the devices and handing them over to us after every work session.
In the months we worked with them, it was possible to see them grow, both outwardly and inwardly. The data we collected only confirmed what we saw happening before our eyes. In two cases, teachers told me how students who had been close to dropping out decided to finish school and go on to university because, as they put it, they didn’t know learning could be fun.
Looking back, I think this has implications for how we introduce AI into classrooms. When students were given meaningful responsibility within a team, particularly responsibility that other students depended upon, many who were quietly drifting through school began to flourish. Responsibility transformed identity in ways that grades alone rarely could. AI may become another opportunity to rethink not simply what students learn, but the roles they are invited to play while learning.
The second aspect I noticed was that students, of any age, were not particularly concerned when things went wrong with the technology, and they were patient in learning how to use a piece of software or hardware so that they got their desired results. In one Grade 6 project, where children were learning about metamorphosis (biology) by creating stop-motion animation films (art), they patiently took twenty shots for each second of their five-minute movie. There was one poor student who had to sit motionless against the trunk of a tree while he metamorphosed into a mound of dirt, moss, and twigs.
Throughout the span of those projects, over all those years, we sometimes heard, “This is hard,” but we never heard, “This is too hard.” Over and over again, I would observe some techie explain how some software worked, amazed at how the other students didn’t balk at the lengthy explanation and endless, “Click here. Then here. Then here. Then here...” Unfortunately, this could not be said of most of the teachers we worked with.
Children often approach unfamiliar technology with curiosity before judgement. They expect to experiment, make mistakes, and gradually improve. That doesn’t mean they should unquestioningly trust everything technology produces. Quite the opposite. It means they are usually willing to persist long enough to discover both its possibilities and its limitations. That willingness to explore without expecting immediate mastery may be one of the most valuable attitudes we can preserve as AI enters education.
Which brings me to the third thing I observed over those years of doing research. The greatest challenge, however, wasn’t technical. It was psychological.
In many instances, it was the teachers who struggled most with learning the technology, not the students. This was at a time when the use of internet technology in schools was only just beginning. When we collected data on the teachers, very few used internet technologies either privately or professionally.
When we approached the teachers about doing projects, they were often curious and open-minded. They were excited for their students. It took them a moment when we told them the students would be teaching each other how to use the technology rather than being taught by the teachers. We always had one-on-one sessions with the teachers as well, which often took much longer than it did for the students to learn.
In my experience, this had to do with two automatic responses that would trip them up. These were, “I’ll never be able to remember this,” when we were giving step-by-step instructions, and, “Why is it so complicated?” when outlining how the students would integrate the technology. This discomfort and internal resistance was prevalent among nearly all the teachers we worked with, regardless of age or the subjects they taught.
In essence, we were devising a situation where students, the ultimate learners, were having to teach each other how to use the technology. At the same time, we were putting teachers, masters at teaching, in the precarious position of being learners.
The shift from teacher to learner was disconcerting. Whereas the students dealt with software glitches or technical failures as part of the learning process, the teachers often felt they were doing something wrong or that the technology itself was inadequate whenever errors occurred.
Our job was often helping teachers see that fixing what goes wrong was part of learning. Instead of expecting everything to work all the time perfectly, we asked them whether they could see technical problems as opportunities for problem-solving. The moment they noticed how unperturbed their students were when things went wrong, and that the students didn’t expect their teacher to fix everything, the teachers themselves began learning how to learn in a way they had not anticipated.
This is not to say the students became the teachers. Rather, by observing their students, the teachers were able to turn inward and reflect upon their own ability—or inability—to deal with uncertainty. This often came with the realisation that much of their professional energy was focused on becoming better teachers, but far less on remaining active learners themselves.
I don’t believe the teachers struggled because they lacked ability. They struggled because they cared. They were accustomed to standing in front of a classroom feeling responsible for having answers. Suddenly they found themselves learning in public alongside their students. That is an uncomfortable position for anyone, regardless of experience.
Generative AI introduces questions our multimedia projects never faced: hallucinated answers, bias, authorship, privacy, and the temptation to substitute generated responses for genuine understanding. Those issues deserve careful attention. Yet they do not remove the central challenge we observed all those years ago. They reinforce it. Adults will once again have to model what it means to be thoughtful learners while guiding young people through unfamiliar territory.
This is a long detour to arrive at the point I want to make. When it comes to AI, leaders, teachers, and parents cannot wait until they feel completely confident before inviting young people to engage with it. Learning and teaching will have to happen simultaneously. We do not have the luxury of figuring everything out and fixing every glitch before students encounter the technology. What we can do is cultivate curiosity, critical thinking, and a willingness to experiment while remaining rigorous in evaluating what AI can and cannot do.
We know AI will change the landscape of learning in the future. What is discomforting to many is that the future is now.
Photo by Markus Spiske on Unsplash
