AI in education: why it's not just about the tools

Why does AI in schools work in one classroom and not another? In this episode of the Hello World podcast, we talk to Dr Mark Peart about looking beyond the tool to the people, beliefs and culture around it, why teacher agency matters, and why schools should ask why before selecting a tool to use.

Transcript

Copy Transcript Copied

Hello World and welcome to the podcast for educators passionate about computing and AI.

I'm Ben Garside, a senior learning manager here at the Raspberry Pi Foundation, and this is the fourth episode to accompany the latest issue of Hello World magazine, which explores critical thinking in the age of AI.

In today's episode, I'm talking to Dr. Mark Peart a lecturer in teacher education at the University of Glasgow.

He is a fellow of the Center for Transformative Change in Schools, and currently serves as program lead for the MSc AI in education.

Before completing a PhD exploring digital citizenship in education for young people.

He was a primary school teacher for ten years in southwestern Spain.

Mark, welcome to the podcast.

Thank you. It's a pleasure to be here. Thank you for having me.

Well, Mark, perhaps before we start, you could start by, introducing yourself and a little bit about what you do.

Yeah. So I'm, a lecturer in teacher education at the University of Glasgow.

I'm currently serving as our, program lead for our MSc in AI in education.

And with the team we're developing quite an quite an interesting stance on what technology in education looks like, what digital education's shaping up to be and how AI plays into that in terms of our pedagogical and critical approach to, to incorporating it into classrooms.

That's really interesting, and I'm looking forward to exploring this more in this podcast.

So the reason you're here is because recently you wrote a piece for Hello World.

And looking at AI through what you call a sociotechnical lens.

So in other words, not just looking at the tool itself, but the people, the value systems, school cultures around it.

What's the main argument that you wanted to make?

Yeah, so the idea really isn't mine. It's from a lot of different authors in the academic literature.

Many of them, are very critical about, digital tools or digital education in general.

Some of them being Buckingham, Selwyn, for example, who I've cited in the article.

They talk about some underlining assumptions that we have when we talk about tech and a lot of the, the edtech, you know, roll out or how that's delivered or pitched to teachers in schools tend to draw on these assumptions quite deeply and don't necessarily fulfill the aims that they intended to do.

So the sociotechnical lens there, what the research suggests about this, is it's not about the tool. It's not about how you employ the tool. It's about a whole range and an array of different elements that facilitate or hinder the process of using tools in classrooms or being digital, let’s say.

So, for example, if a teacher sees that it's a tool is easy to use in the classroom, it aligns with the pedagogical goals, however, it's quite difficult to use or children find it quite difficult to use. That's going to condition whether they pick that up or take that into the classroom or not.

If all of that is positive, for example, but the school policy is quite negative in terms of digital technology, as we're seeing with all of the prohibitions and bans that are currently in the public debate, then that also can condition whether digital rollout or digital education moves forward or actually take steps back.

And I suppose this is, is it new with AI technology or is what you're suggesting this has been a problem that's been around for a while, it's just AI maybe is accelerating or it's the latest version of technology that we're talking about here.

Yeah, I think there's an element of both to that.

So, as a tool itself, so if we’re talking about AI as an app that's assisted with AI or it's a generative AI tool such as ChatGPT or any other.

Then I think the application of that is very similar to previous tech. Because we are trying to feed that into the classroom environment.

We are looking at ways to learn with that or learn about that, if we're looking at AI literacy.

However, there's another element to this which is the other applications that AI has.

Such as detection tools, the surveillance element or monitoring part of that, which leads into other questions and takes us out of the classroom, but still within an educational context.

Yeah. Got it. Okay.

I mean, I can certainly relate as, my experience being a teacher of, you know, some new edtech product being introduced into the classroom.

But can you walk us through, from your experience, or from research, what often happens when that new technology is brought into the school?

And how might different teachers react to that?

So, from experience, and I think also mimicked in what we know from research like Rogers’ work around edtech profiles or adoption profiles.

We always have, or we can think about that colleague who's very skeptical about technology and the colleagues who are very enthusiastic or optimistic about, you know, the tally-ho let’s go with it and see what happens and everyone in between.

And I think all of those points of view are still valid.

And I think part of the intention of the article, but also the work that me and my team are doing at Glasgow is recognizing the whys behind those stances.

If we see a tech optimist, let's say, and a tech skeptic.

Both of them would have founded answers or founded judgments or reasonings for why they feel that way.

Most of it might be linked back to what they believe education to be, what they believe learning to be, how learnings demonstrated from learners, etc..

And then the questions of how does tech marry into that?

Or how do we we knit the two things together or all of these views together?

So I think it's a wider question around the whys there as well, in terms of understanding where people are coming from, what the decisions are.

Because somebody might reject AI and be completely valid to do that based on, you know, philosophical or educational beliefs.

Okay, so I'm going to bring you into a non-AI related example here because, as I was reading your article, I was thinking what technology when I was in the classroom, did get brought in.

And the one that kept on coming back to me was interactive whiteboards.

Now, you might make judgments about my teaching here, but, every classroom in the school that I taught in had an interactive whiteboard.

And I can confidently say that very rarely that I use any interactive features on that interactive whiteboard. It was just a projector, really.

So I suppose what I wrote down was here is, you know, why do you think that happened? I mean, I'm not asking you to question... I mean, by all means, comment on my teaching if you'd like to.

But, you know, have you got any other examples where you've seen, like, technology being introduced that just hasn't worked? And why has that happened?

Well in work that we've done with colleagues, we also equated that to the calculator when that was first introduced.

A lot of, you know, maths teachers etc., will have resisted the calculator being part of the classroom because it takes away, what we then understood the demonstration of learning to be from a learner, in terms of output or process of doing the arithmetics etc.

But going to the example about the interactive whiteboard.

I think it's a really good example to see how it's brought into the classroom, but there's no reflective practice change there.

Or there isn't a change from a reflective practice point of view in terms of how we use that as teachers.

I had them in my classroom as well when I was a primary school teacher, we used a variety of options and but generally it was the projector.

It was the projector screen.

And so I think everybody does it.

And there's moments in time when it's useful for that.

But when we think about why that tool is there, how does that link into my use of education, for example?

So if I think education is a socialising, you know, education has a socialising purpose, then an interactive tool that's basically one person looking at the screen doesn't doesn't foster that purpose or doesn't feed into what I believe education to be.

So that's why the team and I and what we're developing at Glasgow at the moment is, is very much on, let's think about what education means to you, what it means in the context in which you're working, and how does technology play a role in that?

Yeah, that's really interesting. And like you said before, I suppose there's not a right or wrong answer to that.

I mean, I think I was quite fortunate being a computer science teacher and a lab of computers, and I also had the tables....

So I suppose I could create that interactive element in other ways.

And actually the whiteboard didn't serve that purpose.

But you're right, I suppose it really could have done.

So as you explained in your article, these tools don't often operate in a vacuum.

You know, in reality, they're dropped into these messy, complex environments, you know, like the classroom, essentially.

So can you touch upon those different types of teachers or educational technology profiles, and what assumptions are we making about those?

The use of technology or how, how technology is pitched into education from edtech companies from, you know, from policymakers, etc., or, you know, from leadership, etc..

There is a... there tends to be three main areas of assumptions, or what Selwyn identified as three key areas of thought processes or beliefs we have inherent and underlying, let's say, in our subconcious around technology and in the narratives that we create with them.

Which, if we can highlight and overcome, that leads us into looking at it from a more sociotechnical point of view.

So I sort of started with the end with the article and...

Going back to them, so we have the three types of assumptions.

So the solutionist approach is very much that tech is going to fix a problem.

And we've seen that.

Maths attainment is low, let's you know, shoot an app into classrooms to better attainment etc..

And I'm not questioning the apps, they will have some benefit, they will have some use.

But the fact that there are other elements or other things that we can be doing and other ways to measure what it means to be literate or mathematically developed, than just using an app.

And that, that sort of hints of how we need to look at this from a wider systems approach.

The other two approaches are quite similar in nature.

One of them is inherent, as we live in a tech world, inherently tech is good because it's part of the world that we live in.

And we know that not to be true.

With AI tools, for example, we know that there's a lot of hallucinations.

We know that there's a lot of questions around academic integrity, for example, and who's actually doing the learning. Or what are we asking students to demonstrate as learning? And can that be replaced by this tool? I think that's a better question to ask.

And the biases underneath it.

Melissa Warr, for example, did an article around how one word changes the output of ChatGPT or any, you know, large language model, in responding.

So the question was around, a student listens to classical music.

Can you create a study plan or something very similar to that.

Or a lesson plan in accordance to that, and then with another student, they asked, can you do with a student who listens to rap music?

And the responses were completely different, just based on that one word change.

So then Melissa highlights quite a big, bias, let's say, or, how the information that pulls together leans towards what we already have out there in society, which is, classifying in different views and sort of segregating these ideas.

So, going back to the assumptions again, just because digital is part of the world doesn't mean it's inherently good.

We need to question it.

We need to take that critical approach as this edition focuses on.

And the final one lists it as a preparatory nature.

So we need to prepare children for this inherently digital future, which I do absolutely believe in.

But it can't just be the motive or the reason for why we include it in teaching now.

It is part of the aims of what we should do is an education system.

But that doesn't mean that everything has to inherently be digital because they're going to need digital skills.

So I'm going to ask you a really a tough question, who's the ‘we’ in this situation?

Because you know, teachers are, as you mentioned before, really busy, a heavy workload. And, you know, there's so many tools out there to choose between.

Where's the starting point?

So, going to the ‘we’ question, I do drift in and out of these different roles because, once you've been in the classroom you're always a teacher.

I think that’s always the case.

But I think that leads again into that sociotechnical approach.

The ‘we’ is the collective part of the education system.

We all need to have these conversations on different levels, in between levels, so that we understand the realities.

The local authority officer, for example, needs to know what head teachers are mentioning, and they need to know what teachers doing in schools and what the difficulties are, what students do, and how parents actually believe that.

So I think it's a societal issue as well as an education systems leadership, practitioner rollout, let's say, or how we view this.

Because one actor in this system doesn't have the answer or the capacity to change the whole system.

So there I think more conversations or dialogue between the different roles that we play, including teacher educators or teacher trainers at university, play in this, you know, environment or in getting it right.

And I think what the narratives seem to be in this conversation so far is that, you know, one tool is not going to fix all problems either, or if it’s suitable for one environment, it might not necessarily be suitable for another one.

So on that, have you got any examples of where the same AI tool has been interpreted differently by two different teachers?

You know, perhaps one seeing it as a useful support for teaching or learning?

And another seeing it as a risk to effort or fairness or integrity or assessment.

Yeah, And I think an easy example of that is ChatGPT or, you know, Gemini, Copilot, any of these interfaces that generate text, images or others.

You're going to have to teach, and this links again, back to those profiles and how we view the outputs and what it does if a teacher, you know, inputs an essay question, I frequently do, I put my essay question for our final assessment and the courses that I deliver into ChatGPT to have a look at what information is coming up, if it's very similar to what we're delivering at the moment.

Not to try and catch students out.

But that on the one hand, I can see is a positive input because it might provide a different light, it might provide something that I haven't thought about.

Or that the literature’s pulled me away from, let's say.

And that could lead me into either rejecting it or accepting it, or incorporating it.

But on another day that could be absolutely horrific and I could, you know, completely change and decide not to use those tools.

I think specifically with AI tools, gen AI tools sorry, we're using them a lot for answers at the moment.

And that leads into this, the notion of cognitive offloading.

Now I'm playing about with the tool a little bit at the moment with with some students as well, to use it as a questioning tool.

Using it as a teaching assistant to provide you with additional questions for you to unpick the answers as the learner, instead of it giving you the actual answer, it gives you the more guided questions and the more scaffolded questions for you to be able to think about this entirely.

So I ask students to think about, or even me or anyone really, to think about a sociotechnical approach to education.

There's some very basic questions that we would maybe ask ourselves around what does policy that like?

What does practice look like?

But really we need to get to a very critical level with those types of questions.

So by tailoring the tool, using the research that we have around critical, reflective practice, etc., all of these frameworks that we know, using those tools to feed into the machines and based on the need on this evidence group and what these authors are supporting, can you support me by asking them additional questions?

That sounds like a really interesting approach.

And, I suppose a difficulty with that example that you've just given is that it feels like there's some element of choice for educators at some not as well, because this technology is out there and the students are going to use it.

So even if you might be, opposed to the technology, I think there needs to be a recognition that actually young people are going away and using the tool.

So how can they maybe the educator harness that technology.

And I think that leads back again to what's the underlining fear there?

Or what's the underlining, let's say emotional response, but also the behavioral response or the attitudinal response from the educator, because some elements might be and actually, we know this from research, lower levels of digital literacy or digital competence, in your own practice.

So some of these inherently, the stances of complete rejection or to reject the tools from the offshot are actually based on deeper values, attitudes and behaviors.

So, it's about stopping there and going into depth and going why is that?

And it's not to convince the other person of actually uptaking the tools, it’s to understand where they're coming from with it.

And if it is a sense of there's a lack of professional development opportunities, there is a lack of digital, you know, technical or digital competence skills, then that can be fixed and that there is support for that and we can we can look at that.

If it is a case of I reject it because it goes against my educational theory or what I believe education to be.

That is absolutely fine, that teacher has the agency to be able to do that.

Completely and, again, I suppose it goes back to not dismissing that as a wrong opinion.

It can play as completely right.

But also, going back to what you said about communities of practice, it may well be that the example you shared about how you using that is really powerful for educators, right?

Because those more skeptical users see actually other skeptical users finding a way to harness the technology to reach the same kind of, educational outputs.

Okay, So let's make this practical.

You've already done a really good job at this, by the way.

But imagine that you're a school leader that is thinking of introducing a new AI tool into the school.

I think the argument that you're making is that the tool shouldn't be treated as a plug and play solution.

So instead of mandating that all staff should use a tool, immediately, how can I lead staff culture so that they feel empowered rather than forced to integrate it?

I'll answer that with my own experience.

I really enjoyed when I was a classroom teacher that I got asked that question by my head.

Instead of trying to provide a solution, or giving me a solution right from the offshot, it's what do you need?

So having that space and the time in the calendar to be able to sit down with colleagues in that community of practice space or even just as a group of individuals who work in the same space, who dedicate an enormous amount of time to a project, which is the education of a generation, then just having that time to sit down and go, what do you need?

Where are the issues?

What are the problems of practice that you're encountering at the moment?

Tech might naturally be the solution?

Perfect and as again, this goes back to, something that I hear quite a lot where maybe policymakers are saying, you know, they see the kind of the gains that generative AI can bring to education.

They see they can generate lesson plans or do assessments or whatever.

But I think the point you're making and correct me if I'm wrong, is the fact that actually we should be looking at the problem first.

You know what actually, where are the workload issues and will this technology solve it and what are the, you know, pinch points, what are the disadvantages.

And as well with that.

And just to answer that. So I think that not leading those answers to broad strokes themes, let's say, so teachers administrative workload is is overwhelming.

It is as a teacher.

But some of that can't be replaced by AI.

Some of that is sitting down because there's being behavioral issues in the classroom and that in itself could lead off into 20-30 different causes, symptoms or root problems that we need to address further, which is not about the writing of the report.

It's about actually delivering the best educational environment possible or supporting learners with said problems that manifest as behavioral issues, for example.

So I think there's there's other...

AI is always pitched as a quick fix that for certain things.

The fastness as you alluded to, around lesson plan development etc..

Yes, but we still do have to go back over those and be critical about what it's pumping out of the machine, let's say, or what it's churning out, to deliver.

Because maybe that’s not suitable for the learners that we have in front of us.

And that doesn't, you know, it doesn't detract from what teachers do in their day to day lives.

It makes certain things a little bit more easy.

It allows for certain time to reflect and evaluate.

But there's a lot of root causes of symptoms there that, if you ask teachers or go into depth with, what do you mean by administrative burden?

That might, you know, it might mean that we create a non-AI reporting system that feeds into a wider digital ecosystem and there's less reports being floated around classrooms.

It might be something like that.

But we've got to ask the teachers that.

Yeah.

I mean, I remember, I suppose the parts of my job that I would say arguably I didn't enjoy would be marking because it was so time consuming.

However I equally understood it was a really important part of my job to really understand the students.

And if I pass that over to a tool just to give me a number or percentage at the end, I don't think I'll get that same kind of understanding of their progression.

There may well be a tool that does that, but I suppose that's what I want to achieve.

You know, speeding up that marking process, yet still having an understanding of their progression, what they’ve grasped and what they haven't.

And I do sound like a broken record now, but that does lead back to what we believe education to be.

Is it a number? Is it a process? Is it a product?

So we have to think about, well, how do we conceive the what do we mean by education and what is the definition of it for us as educators?

Is it the piece of work that comes out at the end?

Is that process?

If, depending on your answer, your views around assessment, your views around AI tools, your views around where you are important and where you need to focus your attention as a practitioner, it's going to be slightly different.

How can teachers make informed decisions around whether or not it's right for them?

Is it a case of needing training on AI literacy, or is it deeper than that?

I think it needs to go further than AI literacy.

I do believe that AI literacy is important, especially within being critical about AI, you do need to know what's inside the magic box.

How the machines work.

Understand where these biases come from.

Because as soon as you know that, you can find it a lot easier.

So I think AI literacy is one part of this, but we also need to think about other support systems that's needed and rethinking what we have at the moment.

Professional development courses around the technical skills or apps or the how-to knowledge, that's predominant in CPD, for example.

Maybe we need to start looking at that slightly differently and scaling that back to sitting down with teachers for a longer period of time looking at their practice.

And I think part of that is also institutions like my own, universities and with the remit in teacher education, to be in schools, to be with teachers, to know what that looks like.

Because we're essentially their knowledge brokers.

I’ve alluded to research now and that's informed some of the answers I've given you.

If this is a teacher conversation, then we're mixing here the experiential knowledge of a teacher, the research knowledge and that leads to an informed judgment.

That can lead to a wider pedagogical reasoning around what we understand and how to make those decisions for an effective classroom.

And maybe this is interesting that, you know, we don't need to necessarily adopt technology to have any meaningful change.

But it's really interesting that perhaps technology is forcing us to have these kind of conversations and, you know, maybe reassess what education looks like and what's, you know, what's right for our young people growing up in the future.

Absolutely.

And I think part of that question, just to underline the thought again, is it's absolutely fine not to want to use technology in your classroom.

I don't think technology has a place in all classrooms.

Arts, music, there are you know, there are technical tints to that, but it doesn't necessarily need to be.

Outdoor learning, etc.

There can be elements of tech that play a part in that, but it's okay not to want to use it there.

Because that is maybe not be the place for it, or maybe not the time.

Just as long as you know the ‘why’ of why you're doing that.

I think that's the message both of the master's program that we're running, but also the work that we're doing with our work with practitioners across Scotland is to uncover those whys.

Not only the why are you doing that, but why do you think you need to do that?

I'm sure many people listening to this, podcast feel quite refreshed to to hear someone taking that stand point.

Okay, Mark.

So quite often AI becomes synonymous with generative AI tools such as ChatGPT.

As we have referenced.

Have you seen any other, really interesting use cases of AI technologies in the classroom that, you know, isn't ChatGPT or generative AI?

We're seeing this more and more.

There is the confusion or the equation of generative tools with what their capacities are functions in AI could do.

There are cases from around the world actually, that might be positive and slightly more negative, or we can be more critical with certain cases for we see in different countries, AI being used in the classrooms to track eye movements, for example, to look at how children concentrate in different lessons, what their, let’s say, body language or bodily outputs, let’s say, are stating about the lesson to improve teaching practice or to monitor and surveil the students.

So there, we can be very critical and we can analyse that is a case of practice with a lot of different thoughts.

And I think I could give another episode to the podcast actually.

Then we're seeing other initiatives like, one of my colleagues at Glasgow is developing, a teacher's professional development tool, a dialogical tool from the classroom which records the classroom environment.

It records what the classroom dialogues are between students and between the teacher and the student.

And it highlights these, teaching moves, let's say.

So when the teacher responds to the student throughout the lesson and explains the content in a certain way, for example, in instructional and in teacher talk.

It actually analyses that or flags that for teachers, either to sit down in a peer observed situation or in other contexts to actually analyse and think about alternative teaching moves that you can do during that time.

Not to perfect it, but to enhance or to focus more on what that learner was actually asking what it needs.

Very similar to what we do for professional development in peer observation, the video analysis of classroom environments, for example.

But now with the recorded tool in a security environment that doesn't doesn't train a large language model, for example, it stays within house.

So I find that to be very interesting.

And again, we couldn't unpick that in terms of pros and cons, safety issues, ethical concerns, etc., all of these are going to be inherent in every classroom in everyday environments.

So these are things that are coming up.

My own team’s using the work that we've been doing on our reflective practice toolkit, which feeds into what I've been talking about in the article, but also about understanding the ‘whys’ of why teachers make these decisions.

We are moving forward with that to design an AI assisted tool for this, to be a partner with the teacher, to understand the decisions that are made in the classroom and try to clarify that more in the process.

So all of these and many, many, many other ones that we can, that we can see around are starting to pop up.

We have AI assisted tools on our phones already.

I mean Google has it integrated now as well.

So we are using it.

It's a case of we need to decide the when and the when and and what format we're going to be using it, I think as well.

And the research projects that you're working on, the tools that you're exploring, what is your starting point for those?

Was it asking the teachers what they wanted?

Yeah, it started as effectively that.

It started as a conversation between myself, I would say cautiously optimistic tech adopter and an extremely skeptical colleague from the university.

It started as conversations around the faculty, around the school.

And then the questions that we posed to eachother, we started writing them down.

That transformed into a series of questions that we used to critically analyse and evaluate the ethics behind it.

That's then created this master's program that we've been developing, and it's led into, what we have this year, an ERC funded bid which is working with teachers to develop the toolkit that we've developed on the back of that even further.

And another question, around your skeptical colleague, how skeptical is your colleague now?

Still quite skeptical.

I mean, very firm on educational beliefs.

But again, I think that we need to create comfortable spaces to have uncomfortable conversations.

I think that's a real need a and if I surround myself with people who think the same as I do, then we don't learn or we don't question things.

It's sort of an automatic given and we just go with it.

And so we, we purposely meet, we find the time in our diaries to meet and actually academically joust with each other a little bit in terms of, well, have you thought about this? What about that?

So I think a key part there is not to move people from where they are to understand why they're there.

I think we both moved in certain areas, but I don't see it as a fixed position either, It's quite a fluid spectrum that you can move or sway between, depending on the what and the why and in what context, let's say.

But it's very valuable to have those uncomfortable conversations, but in a comfortable space.

Thank you.

Final question, if we were to revisit this conversation in five years time, what do you think success would look like?

You know, what would indicate that schools have moved beyond debates around AI tools and are genuinely engaging with AI as a social, technical issue.

So I think that and a quite a necessary next step or a milestone to move us in that right direction is to move past, where do we use this tool in the classroom?

To move to, as I was saying, the why.

If we move to the why, but we have the conversations as practitioners, as groups and we decide instead of that, or sometimes in certain schools, top down approach to we have this tool, this is a decision from local authority, etc., whatever.

And we start talking and engaging teachers in that discussion.

I think there will be a better uptake because it's actually serving a purpose that teachers have defined.

And that might differ from classroom to classroom, it might differ from school to school.

And that's how education is supposed to be.

So I believe those types of conversations and questions are needed.

And I think that needs to happen on several levels.

And as I alluded to in the earlier part of the podcast, the conversations between these different agents and levels, to be able to understand one another, we tend to do a lot of talking, but not a lot of listening.

And I think that's where we need to draw it back to you, to draw back to what do you need in the classroom?

Teacher agency is is proven to be one of the main, lynchpins, let's say, in what effective excellent educational teaching quality looks like.

Let's get back to that.

But we need to listen to teachers.

That's a really great message to end on.

So thank you very much, Mark.

It's been a pleasure having you on the Hello World podcast.

So thank you for joining us.

Thank you. It's been a pleasure. Thank you.

Well, that's it for today.

If you'd like to read the full article that inspired today's episode, you can find it in the latest issue of Hello World magazine.

And whilst you're at it, if you haven't already, head over to helloworld.cc to subscribe, explore our back issues and listen to more episodes.

What are you waiting for? Until next time. Goodbye!