Can AI reduce teacher workload in Iceland’s schools? A Conversation with Þórdís Jóna Sigurðardóttir

Welcome to the Future Learning Design Podcast.

From our aspect, you know, success is more clarity. It's not speed. And the pilot is designed in a way to give us reliable, nuanced understanding of opportunities as well, you know, the challenges and limitations from the teachers' real experience.

Hi everyone, my name is Tim Logan, and thanks for joining me on the podcast this week.

This is the first part of the 4-part series exploring the ways in which artificial intelligence is impacting the lives of teachers and young people around the world through the lens of Anthropic's recently announced partnerships. In this episode, I talked with Þórdís Jónna Sigurðóttir, the director of the Directorate of Education and School Services in Iceland. In Iceland, who is exploring the implications of AI for teachers' workload and working conditions in partnership with Anthropic, Google, and the Icelandic Teachers' Union. I was struck by how significant the learning focus of this pilot was, with a genuine openness to be both careful and curious in exploring the implications of AI in a country with diverse learning needs and contrasting school contexts, both urban and very rural, in a historically very decentralized system. In the leadership of the Directorate of Education and School Services, which has been active since 2024 and taking over from the previous Directorate of Education, Þórðís Jónna plays a key role in promoting the education system in Iceland and implementing the government's education policy.

So, um, no, thank you very much for taking the time to have the conversation. I'm really, really looking forward to getting to some of the detail of it.

My pleasure, my honor.

Excellent. So perhaps it would be worth just for listeners who haven't thought much about the educational context in Iceland, maybe it's worth just kind of setting the scene a little bit for listeners before we dive into the specifics of the exciting AI work you're doing.

Yes, actually, it's quite an interesting question for me just to, you know, reflect on how would I What will I say about the Icelandic education system? Because in many ways we have a lot of strength in our system and you can see it both from the other measurement in PISA than just from the studies. And for example, we have a strong resilience. We see that children, they are happy in school and they, you know, they feel the belongingness, they have friends. And they trust their teachers. So we score quite high on all these measurements. When it comes to the outcome from PISA research, then we have been declining quite fast. And the last measurement in 2022, we are among the lowest. So obviously there is, there is something changing in the school, and we really don't know what it is. So we trying to do different things to make sure that we will meet it where it is lacking. Because at the same time, we really would like to hold on these strong positive things that we have. And, you know, we have inclusive schools. And of course, for last, what do you say, 10 years, there has been an increasing in school with children with different backgrounds from abroad. So We guess that the, just the workload for teachers is much more than it was.

Yeah, from what you're saying, I think that's very interesting and a common pattern across other countries. I think increasingly the kind of the mix of needs that young people are bringing into the educational spaces. Absolutely. But I put this in later on in the flow of the questions, but maybe it'd be worth just looking at now, if that's okay. There was a, I saw a really interesting note from the OECD about how you were having real success in bringing teachers into the education system because I mean, that's another thing that global systems are struggling with is teacher attrition and, you know, because of the workload and because of the challenges, they're not necessarily staying or coming into the profession. And there was a really interesting success I saw, yeah, in reversing the decline in attracting young people into the training as teachers.

So exactly. And, you know, we have made, you know, deliberately sustainable effort to rebuild the teaching professions. And we, you know, we find the results are visible and we have managed to reverse the declining in teachers graduates. And, you know, after years of falling numbers, but we still have though a lot of work to do. But if you want to go into what we did is that, You know, in general, we have been raising salaries and we think that, you know, of course, is a huge impact into this. And we are trying to, you know, take into account the workload and pay those who have the harder work to meet this extended workload that we talked about. And, you know, at the same time, we are opening the pathway so people with diverse academic background can complete a master's degree in teaching. So it has helped us to have more teachers. And then at the same time, you know, for those who, you know, because it takes 5 years in Iceland to have the license to teach, and we are paying part of the 5th study year to reduce the financial burden. And then what we are working on, and I'm not quite that everyone will agree that has been a fully success, but we are trying to expand the practicum and the work-based learning. Strengthening mentoring for new teachers, etc. Because we have too many teachers still that decide in a few years to do something differently. So we are struggling a little bit with that. So we are trying to increase the mentorship within the system.

Yeah, interesting. I mean, again, it's just an interesting question to me because I think with, you've got on one side, you've got children coming in clearly with diverse needs into a quite a standard, historically standardized model. And then you've got teachers coming to work in that standardized model and not necessarily feeling like it's the most generative, positive place to work. And this is globally, this is not necessarily a reflection on Iceland. And so you've got this, a really interesting kind of dual need to really, I think, radically think about what kinds of spaces are we inviting young people into for this thing called education. And I think, I mean, that's definitely what this podcast is very much about, but it feels, those questions are becoming even more urgent about how we address that. And maybe some of the things we've tried previously aren't necessarily working and we should think some, you know, how do we create some really nurturing relational spaces for the people in them, you know?

Thank you so much. I think, you know, we really need to do better there as well here in Iceland. But yeah, definitely more environment that nurture them, you know, from different aspects. And it is one of the most challenging jobs in the world to be a teacher. So it is, it should be quite obvious that you will not be good at it from day one. So you really need support. And as you know, as you mentioned, it is getting more and more challenging and different from, from what it was.

Yeah.

Because we are expecting much more than before.

Absolutely. I would also though add it's almost one of the most human jobs in the world and we've tried to, we've tried to turn it into not that. We've tried to systematize it. Systematize it and industrialize it and all those things.

So yeah, but at the end of the day, you just, you just sit in front of the children and they need you and you need to lean in. So it is, yeah.

Yeah, no, absolutely. And so then interestingly into the technology and hopefully we'll touch on the way that these different technological arrangements and AI, et cetera, will also help with the workload for teachers and support the teacher working conditions. But yeah, perhaps you could just give, again, like lead us into how did this come about in terms of this, an agreement or partnership with Anthropic and with Google to really step into the space of AI in education. And yeah, how did the thinking evolve to get to that point?

Yeah, maybe I can start by telling you that the institution that I'm director for is just 1.5 years old. And the education system in Iceland has been very decentralized. So now I think that we more or less all agree on that we need more centralization because as most of people know, Iceland is very small but with a big country. So we are very spread around and it is quite expensive and expensive for a very small municipality to have the responsibility for everything and very few people able to fulfill the obligations. And again, talking about how, you know, multiple tasks needs to be addressed in the school. So we are in this new strategy here in Iceland to have it a little bit more centralized, though without, you know, taking everything from them. So when we draw the new strategy, we found very soon that we needed to do it with digitalization. And that's where you can have much more power, much faster and secure way to bring all the messages and all the things that you want to deliver to support the teachers in their job. So we are adding many assessment, different kinds of assessment tools, both more or less all formative assessments. And then, you know, different kinds that support when you see that the the outcome is not good enough, then you have other tests to diagnose what kind of things is going on and how to meet the children. So in all this, we found very soon that AI could help us a lot, both with diagnosing the results from different tests and with— because here in Iceland, we at my institution, we have all the student material as well. So we saw that we could adjust the student material to meet each and every child in a better way. So of course, you hear all this that AI is what we can have to help us with achieving this strategy. And at the end of the day, it is how we can bring the national curriculum towards each and every child in a way that meets them, their, you know, way of learning and how to achieve in the best way. So with having, you know, the national curriculum and all the different tools that we are helping teachers to meet each and every child, we decided that we would like to test how AI could help because we don't know. So we decided to do it very carefully. And then at the same time, we know that teachers, they similar to us, you know, often we are a little bit ashamed of using AI, but as well some people are ashamed of not using AI.

Yeah, sure.

Because the worst thing that you hear is that what you're doing is so AI-ish. Yes. So it's a little bit dilemma, you know, and the same for teachers, you know, of course they would like to find out if and how AI can help them, but at the same time there's a little bit shaming around this and not knowing how to do things. So we decided to do it in cooperation with the teachers' union. And in that way, we knew beforehand that we would have more trust with the teachers because the invitation came from the teachers' union. And with that, they knew that they have their backup. So if something doesn't go what we expected, then they have their union. Behind them. And I think that helped us a lot with gaining the trust that we are not trying to find out how we can, you know, find extra times and use it against them, etc., but really to find out in what way we can use AI to meet each and every child in, in a better way.

Yeah. And so just to clarify, that's then a partnership was initiated with Anthropic and with Google to then be part of that pilot process?

Yeah, we were very clear with them what we wanted to achieve, that we were not going to prove anything or deny anything. We just wanted to make sure that we were learning from teachers, to ask them what their belief is from trying different things. But at the same time, you know, we have webinars and training from both Google and Anthropic to give them ideas about how they can use it, in what way it would be practical. Then at the same time, we have a learning community within the people that are using it and to have a, you know, conversation about what is helpful and what they have been struggling with and ask all the, if you can say so, the good questions and what you might feel, you know, stupid question or whatever, you know, you just have this environment moment that you feel trusted, safe.

Yeah. No, that's great. And I mean, I found it a really interesting dimension as I did the research into it because it was interesting when you read the press release from Anthropic, there was quite a lot of hype, I would say, around what this was. It was a huge, it was the whole country and it was, it was AI everywhere in every school. And then actually when you dig into the detail, you realize It's a pilot, you're learning as you've described, and the teachers' union is a really interesting dimension to that because I mean, you've clearly, as I read the director, Jelmer Evers, say that it was a powerful example of social dialogue in action. So as you're saying, you're engaging with the different stakeholders in a spirit of learning, it sounds like, which I think is a really interesting perspective on it because I think sometimes the perception from the outside is the, the big tech corporate takeover of education. I mean, this is like a money-making opportunity for big tech. And personally, I, you know, I have a lot of reservations about that. I'm sure many other fellow educators do. You know, this is a very human enterprise, as we were saying, you know, how do we create nurturing spaces for young people? That's not about turning them into future customers of Anthropic or Google. Sorry, chaps, but—

Exactly. That's not the passion that drives you to be a teacher.

Precisely. Exactly. So, but it's an interesting one because I think there's a lot of that kind of conversation coming up with the AI hype. So I suppose the question there is what's the process of learning? How will that look? And what kinds of things are you kind of looking out for in terms of the success criteria, let's say, you know, in terms of whether you'll know that this pilot is kind of successful in inverted commas or not, perhaps?

Yeah, thank you for the question. It's a really important one. Because from our aspect, you know, success is more clarity. It's not speed. And the pilot is designed in a way to give us reliable, nuanced understanding of opportunities and as well, you know, the challenges and limitations. And, you know, again, from the teachers' real experience. And then I can say, you know, what we want to understand is, you know, for example, how AI changes workload, you know, does it meaningfully reduce low-value tasks or does it introduce new layers of supervision, checking, and correction? We need to understand this in a deeper way. But, you know, secondly, you know, how AI shapes teaching quality. Does it deepen insight, strengthen teachers', you know, way of doing, or is it a little bit more risk-narrowing creativity? You know, we have seen research that shows that using too much decreases creativity. So we need to understand that, and as well from a pedagogical diversity. We can also mention that we are very, you know, keen with trustworthiness and safety. You know, is it stable, predictable, you know, aligned with data protection, etc.? And the same in trust, you know, for teachers. And one more, you know, thing that we would like to look at as a success, if it helps us with equity, you know, does these tools support teacher equality, you know, equally all, or is it, you know, widening the gaps between schools and districts, you know, we would like to understand that as well.

Yeah.

But then as well, you know, hopefully the success in this pilot project is that we are better prepared how we will approach students. But that's not what we are doing now. But of course, we have started to think about it, you know, how will we do it? Because it's not a question of if, but how.

Yeah, no, absolutely. And I love that you bring up that equity question because that was one of the things that I was curious about was, you know, as you're thinking about representative places from across the whole country, —yeah, I can imagine quite diverse— I've never visited Iceland, I hope to one day, but just quite diverse contexts in which this is happening. So schools with very, very different contexts and situations. And there is always a question of, you know, does one size fit all? Or historically, sometimes education policies have been that. And actually, how do we create a level of opportunity and valuing of the diversity in the system? That allows people to come to it in the ways that they need to, that suits them more. And of course, there's the access question of reliable internet access or reliable device access and those kinds of, you know, very practical things, right?

Yeah. You know, this is one of our biggest concerns today is that there is going to be even more gap between— because in Reykjavík we have one-third of all the children. And if you look at the Reykjavík area, we are more or less with two-thirds of all children in Iceland. So in some of the municipalities, there are very few, and we hope to meet them even better, but we don't know how or if it's going to be true. So this is one of the things that we really would like to understand and how we can then support teachers in the smaller areas to use it in a way that we are sure is going to help.

No, that's amazing. And there's the other topic that I'm very interested in around this question of, of AI and I mean digital technology more generally, social media, et cetera, is the wellbeing question for young people. 'Cause I think there is a clearly an efficiency of personalized learning and bringing devices to young people and, you know, allows them to access learning in a way that suits them in some ways. But there's also the kind of question, of how much time are they spending on screens, how much time are they actually outside in nature as a, you know, a counterpoint to that. I mean, it, it, that how much attachment are they developing, particularly with as AI chatbots become more kind of responsive, you know, to people talking about the hacking, not only hacking attention, but hacking attachment. So young people becoming attached in some emotional relationship with chatbots? I mean, these are, these are quite scary questions, right? So how are you holding some of those questions in, in this pilot?

Yeah, in this pilot we are not, but definitely, you know, this is what we are just starting to look into. And as you mentioned, it's quite scary, but at the same time, you know, we as well see opportunities as many others, and we have, for example, seen one research showing that they are having better outcome in some of the tests, but less education. So that's what we make. So, so probably teachers need to test differently to ask different questions so that you make sure that you are not just using AI to do what you're doing, because that doesn't prove anything except that you can ask AI the same question that teacher ask, ask me. So we need to make sure that education is what we are, you know, aiming for. And therefore, part of what we are doing is working with OECD to make sure that the competencies in the national curriculum is aligned to what we are doing with AI for teachers, because we have agreed on the competencies that we want to help children to achieve in their education and make sure that they are still there. Yeah. That's not going to change, but then there might be a change how you will achieve it with children. But then, you know, the, a good school is a school that achieve to deliver their competencies. So we are still working on that and trying to make sure that AI is trained in a way to assist teachers to achieve the competencies.

Ah, very interesting. And I've, I've actually just had Michael Stevenson on the podcast, who was the primary author on the OECD Education for Human Flourishing conceptual framework. And so, so that, you know, we talked a lot about how competencies around, yeah, how do we think about competencies of young people in relation to this broader purpose of flourishing? And that, again, that's a big question when you bring in AI and how does, you know, does AI support young people to flourish? And teachers to flourish in the workplace? I don't know, it's a big question.

Exactly, and you, you cannot lose sight of this. And as well, just for the general competencies within the all subjects, even if it's literacy or math or whatever, you need to make sure that the competencies is there in right way, you can say.

Yeah, there's an interesting narrative as well I find fascinating around, it's a kind of a progress narrative of the inevitability of AI progress. I mean, it's, you know, it's almost, as you said, there's a lot of shame in using it, but also a lot of shame in not using it because there's this perception of some kind of inevitable wave and we all basically just have to kind of surf the wave and keep moving in that direction. It's interesting because it makes it harder to ask negative or critical questions, I think, with the feeling of inevitability. And so, What I really appreciate about what you're saying is this kind of learning stance around it. And I think if only we could hold more of that kind of a stance. So it's critical, it's questioning, it's reflective, but it's also open to learning. I think that helps us because I think, I mean, for example, recently I've also had another author on the podcast, Jared Koenig Horvath, who's just written a book called The Digital Delusion. Which is effectively saying that digitalization has really damaged cognitive abilities and broader competencies in young people. So there are, you know, there are many sides to the discussion and it sometimes is difficult with that feeling of the inevitable progress of AI and it's just gonna, it's just gonna come and take all our jobs in the kind of the hyperbole that, that gets, you know, in the tabloids, et cetera. But it, you know, there is a feeling that brings that shame of not, for teachers, of not getting on board. So I don't know. I mean, I don't know if any of that is coming up for you as you think about this.

Definitely. I think that was, you know, one of the big reasons that the teachers' union wanted to enjoy us, is that to help teachers to step forward, to try and to not to be scared and to make sure that we make this trustfully and safe space for them. To find out, you know, is it something that is adding value or, and in then what way is it adding value or is it, you know, maybe something that we should not use at all. It's probably not going to be the case, but at least we allow ourselves to ask. And, you know, I think in Iceland, as in most countries, the teachers are getting older. You know, it is not a sustainable curve. Maybe just we are trying to, you know, bring a younger one. As I said before, they're getting a little bit more, but, you know, and for the older teachers, it can be a certain threat or uncertainty how I use it, or if it's too complicated, if you need to change something, etc. So probably the younger one using it more. Aggressively even, and not knowing if it is secure. And then the other not wanting to try because they're just maybe afraid or not knowing how to do it. So I think we have both groups and yeah, and we needed to combine them, learn from each other because you need to be careful and at the same time you need to be curious.

Yeah, it's interesting, isn't it? It's almost like a microcosm of just any social change. I mean, you will get older generations who perhaps are more resistant to the change. But also are holding on to a lot of wisdom of the ways that are important and valuable and should be held on to. And then the new, the younger people coming in with different ways of doing and being. And it's a fascinating thing. But as you say, I just really like that if you can hold a space for that in a trusted way for people to genuinely be able to feel that their perspective on it is valued. And I just think it's a great move having the teachers union as an arbiter for some of that. I think that's, That's really, really interesting. So yeah.

And at the same time, you know, we have a little bit less than 10% of teachers participating. And, you know, the theory about the tipping point, you know, when things getting that popular that they take off, it's around 12 to 30, 40% that you need things to go up to, to change for good. So we are probably, you know, somewhere around that, you know, you know, we have that many people participating, that it is quite obvious. And we have already started to think that when this ends in beginning of April, that what will we do, you know, what, how will we continue, and what, even though we don't know and we don't need to know exactly, but at least we need, I think we are quite sure that we will need to continue. But in what way is it going to be in a teaching, training, mentoring, you know, probably need to have a combination of all this. And then we are to work in much more cooperation with the municipalities and the universities as well.

Yeah. Okay. Yeah, it was interesting. I mean, that was one of my questions just in relation to practically when might you have some, because it'd be interesting to hear, you know, maybe invite you back and hear the reflection on the other side of once you get some more data. But the other question I had was, are you, how are you involving young people in the discussion or the reflection on the process? Because obviously this is, from what I'm understanding, it's more about the impact on teachers, the AI to support the teachers' roles. But is there a space for the young people to also bring their voices in to kind of reflect on the pilot as well?

Yeah, we are struggling a little bit about how we take that on. I think we need to do it in a very good cooperation with teachers, and we would like to learn from other nations at the same time, you know, what would be the best way, because it is very sensitive, you know, because children only have one childhood. You know, you cannot risk it with some kind of a pilot project, you know. So it is, it is a big challenge, I have to say. And maybe we will start with the older one, upper secondary school, maybe 16 to 18. Yeah, that might be a good one to start with, and then we might need to, you know, go a little bit down. But I'm sure that, and I know that, there are teachers who are using it very wisely and in a very pedagogical way that we can approve on. But we haven't framed it in a way that we feel good yet.

No, it's interesting. I hear you very much. It's a challenging question in terms of how do have the care and the responsibility for the young people and not just— I mean, when I was a senior leader in a school in Berlin, I remember a parent saying that, you know, they didn't want their child to be part of the experiment. Exactly. And there is this challenge around it. And at the same time, I was reflecting on that recently that actually we've kind of let our young people be part of experiments with somewhat unregulated social media in their lives and access to smartphones, etc. I'm not sure what it's like in Iceland, but, you know, I have 3 adolescent children and they have smartphones and, you know, I worry about what I've allowed them to have as a parent. So it's quite interesting. They have been part of an experiment and it's really interesting to know how to hold that.

Yeah, I think that we all are a little bit worried about that now. You know, it seems that it has more effect than we would like to know. So maybe we would like to be a little bit more cautious about this because I use AI a lot and I feel I get a little bit addicted to it. Yeah. I've found it so securing to, you know, let it, you know, reframe what I'm doing, what I'm thinking, you know. Maybe it's just when we start to use it that, you know, use it a lot and then we find that, okay, it doesn't help in all situations, but But it seems that it is, you know, you cannot use it as a handicap, you know, you need to use it in a structured, creative way. So, and know when it is right or wrong.

Yeah, it's a real challenge, isn't it? Because it comes back to where we started in terms of creating these kind of nurturing spaces for young people. And if they have secure attachment and they feel they're in a safe, trusting space that's nurturing them, they're probably more grounded to be able to make good, in inverted commas, choices around the technology use, right? Whereas not everybody has the privilege of that kind of context. And so that feels like an important piece of this as well. Just kind of coming full circle back to where we started. Lovely. Well, thank you. I mean, this is fascinating and it'd be so interesting to track your learning and, you know, report back as you get more information. But thank you so much. It's so interesting to hear about it.

And thank you so much. It was really inspiring talking to you. Thank you.

We hope you've enjoyed this episode. Please feel free to continue the dialogues with our guests, with us at Good Impact Labs, on our blog or on social media, or within your own networks.