Via Dan Willingham, I discovered this new report by the American Psychological Association on children, adolescents, and educational technology. Dan was one of the authors. Much of what is in the report will not surprise regular readers of this blog. Take e.g. the finding that engagement is not the same as learning. This reminded me of Robert Coe’s well-known list of poor proxies for learning. A pupil or student may do well inside an application but remember almost nothing of that experience once the application is closed. More versus less screen time is not very meaningful on its own, even though screen time can cause quite a bit of (moral) panic. What is far more important is what the individual is doing, when they are doing it, with whom, and especially what that screen time is taking time away from.

Still, I think this is a pretty strong report. Mainly because the authors avoid one of the classic traps in education. That trap is pretending that only one outcome matters.

You can see this, for example, in the way the authors of the report approach screen time. According to them, the question is not simply how much time children or adolescents spend in front of a screen. What matters more is what has to make way for that screen time (check also these earlier posts). The report explicitly points to possible tensions between technology use and sleep, social interaction and direct experiences. This is particularly true for younger children and adolescents. All of these can make a difference to a child’s development and wellbeing, with both direct and indirect consequences for learning.

I think this matters because discussions about technology, children and adolescents, and education have a remarkable tendency to narrow or become polarised. Think, for example, of the debate surrounding Jonathan Haidt and his critics. Sometimes these discussions focus mainly on learning gains, with everything else treated as secondary. At other times, the focus is on wellbeing and development. In this case, learning suddenly seems to become a slightly uncomfortable topic. One of the strengths of this report is that it tries to take a broader view.

The APA sticks quite firmly to a clear definition of learning: students should not merely demonstrate knowledge, skills or attitudes while using a tool, but should also be able to retain, understand and apply them later. Interest, motivation, attention and participation can support this. However, they are not in themselves evidence that learning has taken place. This is a useful corrective to EdTech companies that like to point to usage statistics, engagement, badges, streaks and happy students. Lots of use is not evidence of lots of learning. Back to Robert Coe.

But at the same time, the report does not say: fine, then let’s only look at academic performance. Relationships matter in education too. So do development, sleep, privacy and accessibility. And whether some groups of students benefit less, or may even be disadvantaged, certainly matters too.

You see the same broader perspective when the report turns to AI. The APA experts are clearly cautious. Generative AI can improve performance without students necessarily learning more. Students may produce better answers but subsequently know less or be less able to solve a similar task independently. Hopefully, this is an insight that has become common knowledge by now.

One interesting practical recommendation follows from this: preferably, do not let AI immediately produce the first answer, first idea or first draft. Let students try something themselves first, and only then use AI to compare, improve or ask for feedback.

But again, the discussion does not stop at learning. For younger children, there are questions about human-like AI, trust and dependence. For adolescents, privacy, independent thinking and emotional dependence also enter the picture.

Something else I think is positive is that teachers and other adults certainly do not disappear into the background in this report. Again, logically enough, this is in line with other research. Technology can work better when adults help students focus, ask questions, provide explanations, and connect the digital activity to what students already know or have experienced elsewhere. The APA explicitly warns against seeing technology as a replacement for human relationships and professional judgement.

That too may seem obvious, but… Over the past few years, we have heard often enough that technology would become personalised, adaptive and smart enough to take over more and more of that role. Quod non.

The report ends with a set of ten recommendations, which I have highlighted in the picture above. It is not an eye-opening set of recommendations, but perhaps that is what makes it useful. We are usually under the impression that it is difficult to keep up because technology moves so fast. But perhaps we should also remember this set of questions:

• What is a student actually supposed to learn here?
• Is the student really learning it?
• What is this technology replacing?
• Who benefits from it, and who does not?
• And what does all of this do to a child or adolescent, not only as a student, but also in terms of development, relationships and wellbeing?

The first few questions made me suspect that Dylan Wiliam had also been involved, but no. He came up with a similar list some time ago. Both his list and this new one can help us have a much richer discussion. This is much better than endlessly arguing about “technology: yes or no”.