Hello and welcome to this OECD webinar with me, Duncan Crawford. Today we're talking about how AI and how to help teachers use artificial intelligence effectively in their lessons. In many countries there are currently few guidelines or rules about the use of AI in classrooms.
So what AI tools and abilities should teachers be using, if indeed they should be using it at all? Well, to join us on the webinar today to discuss all of this, I have a fantastic group of panelists. We're going to share their thoughts on how teachers can use AI to improve learning outcomes.
Joining us on the panel today, Punya Mishra, Director of Innovative Learning Futures at the Learning Engineering Institute at Arizona State University. Masalmati Sharma, founder and president of Nonprofit Workshop for Me, which is based in Luxembourg, and Samo Varsic, an OECD policy analyst on the education for inclusive Societies project. Thanks to all our panellists for joining us today.
To kick things off, let me hand over the webinar range to Samu Vasek, who's going to give an overview of some of the latest OECD research in this area. Over to you, Samu. Thank you very much, Duncan, for this introduction. And I just want to mention that this presentation is stemming from a working paper, which we are currently drafting. The title of the working paper is the potential impact of artificial intelligence on equity and inclusion in education and it will be published in the coming days so stay tuned for that.
Now the motivation for the working paper were basically two questions and I think these questions are also relevant for this webinar. First, what are the AI tools that are making an increasing presence in our classrooms and secondarily what should policymakers keep in mind in regard to the impact AI tools might have on equity and inclusion in education? In order to answer the first question, we have adopted a framework developed by Wayne Holmes and his co-authors and basically categorized AI tools into three groups. The first groups are learner-centered tools to support equity and inclusion, and these are designed to enhance the learning experience of students.
The second group is our teacher-led tools to support equity and inclusion and these can assist teachers in their instructional as well as administrative roles. And then finally there are other institutional tools that can foster equity and inclusion and these are aimed at addressing broader institutional objectives such as identifying students who are at risk, assisting with admissions and etc. Now because of the topic of this webinar I'll be focusing mostly on teacher-led tools but I imagine we will cover or touch upon the other types of AI tools as well.
And in order to describe the opportunities and challenges that these teacher-led tools bring with them we have adopted an approach where we describe the types of tools and thereby describe the opportunities as well. And so starting with AI-powered robots these are making an increasing presence in our classrooms. Teachers often sometimes use these to support students with autism spectrum disorders to help develop their social skills in a non-threatening and engaging environment. Secondarily, there are AI tools out there which can help teachers to curate learning materials. This concerns not only the development.
of the materials, but also translation to languages for students whose language is not the language of instructions, for instance students with an immigrant background. But it can also help teachers and lecturers to provide live captions for students, for instance with hearing disabilities. Thirdly, you know, there is a plethora of tools out there which can help teachers with grading, although I should warn teachers that these often there is research out there research studies out there which also highlight that these can often perpetuate biases and and biases in in in terms of grading then fourthly there are ai tools which can aid teachers in the identification of some student needs and special education needs for instance based on audio records or writing records ai tools can help teachers identify dyslexia and dysgraphia of students with an impressive amount of precision.
And then finally, AI coaches are making an increasing presence in this domain. And this can help teachers to provide feedback on their lesson plans, for instance, but also to broaden their knowledge, their content knowledge for a particular topic. Now, while all of this sounds pretty great, I also want to highlight a few challenges which these tools come with.
And then firstly of all, I want to highlight the costs. Now, you can imagine that AI-powered robots, for instance, can be prohibitively expensive. Although, we should also say here that AI tools have also been highlighted as improving access, and many of them are indeed free. Secondarily, I want to highlight that there is a balance to strike between the objectives we have in our education systems in terms of delivering knowledge, skills, and developing competencies of students for the 21st century, but also equity objectives and the objectives of companies, of what are often profit-seeking companies, which develop these tools. Thirdly, and this is an extremely crucial point, we need to keep in mind that educators need to be equipped with the knowledge.
on which tools are appropriate to be used in which context. And then finally, there are many other challenges which I won't go into details, but I imagine we will touch upon these in the following discussions. And these pertain to bias, privacy and data security issues, or accountability. You know, who is responsible if an AI tool provides a wrong or incorrect or misleading answer? Now that's just a very brief overview from me and back to you Duncan looking forward to the discussion.
Thanks so much for that presentation Samo, very interesting indeed and thanks for everyone who has joined us on this webinar. If you do want to get involved then please use the chat function at the bottom of your screens. You can type in your questions and comments and I will try to get through as many of those as possible. Now though let's get the discussion going and to kick things off, let's start by talking about what AI tools can actually do to benefit teaching. Punya Mishra from the Learning Engineering Institute at Arizona State University, why don't we start with you?
What types of AI tools do teachers at your university use and for what purposes? Thank you, thank you Duncan and thank you. Sama for that introduction.
So glad to be here with everybody this early morning from Phoenix, Arizona. So I think one of the things to understand about this technology that it's a very multifaceted technology. And that so there are many possible uses, some of which sort of Sama introduced to us in the beginning.
My video seems to be out of focus. Let's see. All right. That's much better.
Thank you. So I think what we are seeing at our university and what our university's approach has been, rather than mandate or require that people use it in a certain way, it has recognized that different units within the university, faculty at different colleges, are going to be using it in very different ways. Arizona State was sort of at the forefront of actually forming one of the first, was in fact the first university to form a partnership with OpenAI, because one of the challenges, like Samo mentioned, is our issues around security.
around issue of data privacy, particularly those information about students and so on. And so what we have is essentially a walled garden where our faculty and students and staff can work with ChatGPT and other tools, but that data would never be used for training their models, that the data would be secure within sort of the ASU enterprise. So I think we have been very sort of cognizant of that.
At the same time, I think what we've tried to do is be very open and flexible with respect to the kinds of uses people would, you know, would have with these sort of technology. So speaking just of my college, I'm in the College of Teacher Education. And so we have more than 50 different faculty members, staff members who are using it in a variety of ways for pedagogical purposes, for research purposes. So, for instance, we have teams of people who are using it for qualitative research. So think about when you do qualitative research, you do inter-rater reliability.
Here we have AI being one of the evaluators that is being now compared with. how humans are using it. We are using it to develop instructional tools in our learning design technology program.
For instance, we have AI mentors who you can ask questions to. Now, are they always right? No, not necessarily.
I mean, I think one of the important pieces that we've looked at is what are the limitations of these tools that how we shouldn't see it as being like a mentor or a truth teller, but rather as a cognitive partner, as a pedagogical partner. So I think how we approach... This technology, I think, is incredibly important. And I can dig into that a little bit more, but I know we have a lot of ground to cover. So I'll pause here and allow you to move on to the others in the panel as well.
Thank you so much for those opening thoughts. Very interesting indeed. Mahamati Sharma from Workshop for Me, let me bring you in now. I wonder how much you agree or disagree with what was just said there, because obviously you're passionate.
about bringing coding to children in particular and mentoring girls and women into the tech industry. So what tools do you think are of most benefit, AI tools? Yeah, so hi everyone.
First of all, a very warm welcome. I can see people joining from really all parts of the world. This is amazing to see that everyone is interested in knowing more about AI.
And I'm expecting that you would like to get practical tips and tools which you can take back and... apply in your classrooms. I see a lot of teachers and you've mentioned with school and region you come from. So I'll start by saying that apart from running a non-profit organization in Luxembourg where we teach kids coding, I'm also the EU Code Week ambassador in Luxembourg and the EU Robotics National Coordinator.
The reason I mention all of this is because using all of these different communities, the idea is to be able to provide practical tips and tools for teachers and students to be able to leverage AI. So recently we put together an AI squad, actually the EU Commission formed it, with actors from different parts, including non-profit organizations, teachers, policy makers, and we came up with an AI report, which is aimed directly at teachers. It has seven different reports, I won't go into the details of that, but in terms of the practical tip that teachers can use, basically depending on their digital competencies, they can actually decide to use AI.
in three different ways. The first one is teaching about AI. So this would be the developer perspective.
How can I leverage AI? How can I teach my students how to classify things, how to use AI for making their jobs easier? So that is one part of AI.
The second one is for AI. This is being able to create things within the classroom and outside in a world that we live in, which is surrounded by AI, where we are able to actually figure out is this information really true or is this actually fake information? Is this image AI generated or is this actual photo of a person? So basically from the user perspective.
And the third one, which Samu has actually gone into the depths of, is using tools that AI provides for teachers to make their jobs easier in terms of assessment, in terms of personalized learning, in terms of making it easy for them to concentrate on providing individual feedback and the things which teachers need to do as part of their job. can be made easier, the administrative parts of the job, by the AI tools itself. So there are really a lot of different ways in which AI can be used.
And I'm sure we'll talk about specific tools as well as we go along. So feel free to shoot your questions. I'm watching the chat, everyone, and I'm happy to answer anything you have. Thank you so much for that. On the question of specifics, let's bring in Samo Varsik.
I don't know if he can answer this or not, but are there specific types of AI, generative AI or companies that you think are doing a particularly good job and are helpful for teachers? Well, there are many companies which I think are particularly helpful, but it is ultimately up on the teachers to decide which AI tools they need to use and which contexts are most appropriate for the AI tools to be used. And we actually have very little internationally compelling comparable research, which would be able to tell us at scale whether and which tools teachers are actually using and integrating into our classroom. Now, this is partially stemming, of course, from the fact that these technologies are relatively new and it takes time to develop an internationally comparable survey and to analyze the results as well. But we do have some evidence from individual education systems.
So, for instance, in England, About four out of ten teachers have used AI tools, and they are talking specifically about language generating tools, to help them with schoolwork. But at the same time, about half of teachers, five out of ten teachers, reported that they've never used AI tools, either for their personal or professional purposes. Now, this survey was conducted in November 2023. As you know, this field is constantly evolving and the figures might look differently today.
But just to highlight that at the moment, we actually do not know that much at scale, and particularly at an internationally comparable level, which tools and how teachers are using them. Thank you so much for that, Sammo. There are lots of questions and comments coming in on the chat.
One comment from Tasneem Slate-Waller, an assistant teacher. She writes, AI doesn't allow student creativity to be explored. Students become dependent on AI and don't use their brains.
Also, another question here, who I'm afraid I've actually lost the name because the chat's moved on, but the question said, is there evidence that AI hinders thinking skills of students? Punya Mishra, let's bring you in here. Are most teachers using AI effectively at the moment, or can it hinder learning? So I think what AI has done is sort of laid bare some of the sort of harder issues. questions that as educators, we should have been asking all along that how are our assessments genuinely measuring what students are learning by allowing, you know, an AI to give you an immediate answer to something which may be actually not very good, but is possibly good.
Does it get in the way of learning? I think absolutely. However, so that means for us, the challenge then becomes as educators is can we be using it in ways that are productive?
and helpful. Coming to the question of that AI actually hinders creativity, I think if you see AI the right way, that it actually promotes creativity. I think Madhumalati can speak a lot more about it in terms of coding and things that way. Personally, I don't know how to code in Python.
I actually have written working computer programs now, simulations in science and math that I could have never done. I have done it with AI. So it has actually improved and and, you know, broadened my creativity, and we are seeing that as well.
But I think the most important thing in this issue is to think about the technology the right way. I think very early on, the technology, when it first came out, people thought it was an advanced form of a search engine that you can ask it a question and it'll give you an answer. It's not that at all. It really is generative. In the sense, every time you ask it a question, it's actually building the answer out of whole cloth, which is what explains why we get so many hallucinations.
and why we get so much variability. So the challenge there for educators is if you're going to be using it, like Samo said, for student assessment and so on, we have quite a bit of strong research showing that you get inconsistent results, which bring in implicit biases and all of those things emerge. Also, be sensitive to the fact that where is the data coming from that these models have been trained on? These data, as we say, are weird, which means they're mostly coming from Western educated, industrialized, rich and democratic countries. And so when I've tried to do translations with it, it does fantastically well with, let's say, Spanish or French, not so good with Bengali, not so good with some other languages, right?
And so I think recognizing the limitations of this technology, seeing it not as an answer, you know, a tool that gives you the right answers, but a tool that you can engage in a dialogue with. I think that's something that we really haven't explored much. And I think if you see it in an educational context, for it to be a partner with you, actually, I would argue, enhances your creativity, enhances your thinking skills, because now you have to critique the response that it has given rather than take it at face value.
So again, coming back to, I think, some of the things that Madhumalti talked about, which is the kind of literacies that we need to develop, the ways of thinking we need to develop around these technologies. You know, at some level, it's not anything new. I mean, this is...
People shouldn't believe anything that I say or anything on this panel says. They should listen to us and then filter it through their experience, through their knowledge. Check what we are saying.
It's the same thing, mindset that we have to bring to this new technology rather than assuming it to be a neutral tool. Thank you, Punya, for that. He mentioned their limitations and challenges.
And there is a question here from Sara Frutos, a teacher at Universidad de Tela. I think that's in Argentina. who asked what is the most useful way to use AI in the classroom, given some of the challenges and limitations outlined there. Madhulmati, how do you go about using it in the most useful way?
It depends on the age of the children and the level of digital competency that the students have. So just as an example, if you have a group of students and you want to introduce them to, you know, what is AI, what does AI do? Then there are very simple tools which help you understand what classification is.
So it's like a little game where it shows you fish and it shows you garbage and it asks you to classify what is fish, what is garbage. you put it into the two boxes. So not only do you learn about recycling and the effects of climate change and how we can make it better, but you also learn about the classification mechanism to do it. The other tools which you can use are like quick draw or auto draw, where there is a word given and you're expected to draw it.
And then you'll see, this is developed by Google, and you'll see that over time, the pictures get better and better because now there is more data of how different people have drawn, you know, whether it is a teacup or a flag or whatever. So this is one of the ways you can introduce it by being a user using a tool which already exists and about creativity for example if you have a particular thought in mind as to you know I need an image which should be belonging to that you know that era of time which should show a student studying in you know this age from this country a future scope or whatever You can give examples like this to Dally and tools like this, and they're going to be able to generate images for you. So these are some of the ways in which you can use AI as a user. Then if your students already have some level of knowledge about digital competencies and all of that, you can go ahead and start creating a chatbot with them. They will understand what it takes to make the chatbot, which is a tool that you're creating, understand the question, break it down into the parts and give a meaningful answer.
So once they go through that, they will learn the challenges that the software is facing, their AI is facing, and learn how to overcome it by actually knowing how to code, how to give the correct prompts and all of that. To be able to answer the question about whether it is good or bad, does it hinder creativity or doesn't, I would like to give the example of a knife. A knife can be used to cut vegetables and make your dinner. The same knife can also be used to hurt somebody. Does it make a knife good or bad?
No, it's the use you're putting it to. The same goes for AI. That's what I would let teachers take back with them. Very, very good analogy there.
Samo Vasik, let's bring you in. It's a slightly different topic, but a question from Kathleen Foley from Universidad Popular Tomolinos. She asks, student safety, including how personal information is treated, is paramount.
How can we ensure that... personal information is protected? I know it's a bit of a side issue to effectiveness, but it is an important question, Simon.
Well, it's absolutely, it's a crucial question in this domain, I feel. And, you know, there are many tools out there, which I think you should be very cautious about once they start asking you sensitive or private information, either about yourself as a teacher, but also, but even more worryingly, perhaps about the students you teach. So I think this could be one of the flags which, I mean, you know, I'm not saying you should immediately disregard the AI tool, but this could be one flag which you could consider as a rule of thumb or as one question which could guide you in whether the AI tool can be used or should be used or should not. There are also, you know, you have your own school regulations, I imagine, which which can guide you in the way how you use AI tools and in the way how you protect student privacy.
And of course, then we have national legislation, such as in the EU, it's GDPR and many other legislations, AI legislation now coming in, which should also guide the development and utilization of AI tools more broadly. Thanks for that, Samo. Punya Mishra, let me bring you back in on the discussion about...
how to use AI effectively in learning. At the Learning Engineering Institute where you're based, you routinely collect learning data, lots of it, and you analyse it to better understand learning outcomes and to improve the overall teaching. What sort of things have the data shown you that you didn't expect? So...
The generative AI work is pretty new at the Learning Engineering Institute, to be honest, given the newness of the technology. But let me just give an example of how data can be used, machine learning algorithms, which are a form of AI, if you look at it, to identify challenges that students face. So for instance, we have a beginning calculus math class, which is very predictive of...
whether you succeed, you know, down the road in college. And we have been able to, looking at the data, pinpoint exact assignments where students stumble the most. And then we can target our sort of support to students at that point in time.
If you don't show up for a couple of classes, or you've missed some homework, or you've scored low on some tests, we can actually target support, which then has a ripple effect down the road. So that's just one example of the many ways that we could use data. I want to just come back to something that Samu said, which about giving information to this AI about your school or about yourself or about your students, what we are finding is these technologies are so sensitive.
So you don't even have to explicitly say the race of the student or the gender of the student. We have had studies where we are showing that just including a word in a student essay. So for instance, if a student has written an essay about how they're preparing for an exam. Like I studied this way, I make my notes, I listen to music. And the only thing in that paragraph is the use of the word rap versus classical in terms of what music they listen to.
We are getting AI and this is not just one model. We are testing it across models. We are getting differences in scores based on an implicit bias that a student listening to rap is less cognitively capable or what have you than one listening to classical music.
So I think the level of sensitivity that we need to. bring to this. This is a really powerful technology. Let's not ignore that part.
It allows us to do things that individually or collectively, we may not have been capable of doing as efficiently fast, but also capabilities that we may not have had before, right? At the same time, also recognizing that it is a dark mirror to us, that all of our strengths and weaknesses just come back reflected. And so building that critical literacy, the building those critical skills about how do we judge information? How do we determine whether it's true or not?
How do we determine, you know, those kinds of skills become incredibly important in this space, particularly as we look at the broader geopolitical spectrum within which education exists. Thank you, Punya. I think many people listening to the example you just gave will be shocked, even horrified that the level of bias there that could happen and the dangers that might... that that could cause. And I wonder about how Mati Sharma, I mean, there'll be some who would argue that, you know, students schools are being used as guinea pigs for this technology, and that it shouldn't be.
What do you say to those who would argue that? So, yes, we should definitely be aware of the biases that exist. And only when those get uncovered will we work to make it happen.
Just as an example, you're possibly aware of the case of a face recognition software. that was not able to recognize a black face when put in front of the app because the data didn't contain enough black faces for it to be able to recognize it. So this actually goes to show that we need diversity even in the development of AI to prevent biases like this.
Unless we take a good population of diverse kind of people across age ranges, across ethnicities, We are not going to be able to make our AI better. So this is one way, you know, where we are trying to push and make sure that the development is going to ensure inclusivity. The other aspect is that we need the regulators to be able to, in what exists today, say this is good enough or this is not going to work and classify it as a form of risk.
So the AI Act, which the European Commission has been able to bring forward like a legal framework, and this is a recent development. is a great start for other countries around the world to also be able to think of what should we get considered for AI. I know that there is a challenge because there aren't enough people maybe at the policy level who are aware of, you know, what the challenges and risks could be. Everybody is going through this steep learning curve and it is evolving as we go. But having frameworks like that in place, having acts like this being thought about and worked upon, and then...
quickly doing iterations on it to keep up with the actual developments of AI is a great way to make it happen. Thank you for that. Sameer Vasik from the OECD, let's bring you back in because given some of the issues raised here about bias and so on, I suppose one of the ways of preventing or trying to mitigate that involves teacher training. And are teachers in general at the moment receiving training about how to handle AI to make them aware of these issues or in general are they not? This is a fantastic question.
And again, I'm afraid I'll have to say that at an internationally comparable level, we do not know. Because AI is a new technology and as such, the data simply has not been gathered yet. Now, we do have indications, however, more broadly about how teachers feel prepared about integrating digital technologies into their teaching. And so, for instance...
And all the piece of evidence from the Thales 2018 survey says that about two out of ten teachers feel a high need of professional development in integrating digital technologies into teaching. But you have great heterogeneity both within and between countries. So for instance, in England, only 5% teachers felt that they have a high need, but in Japan, 40% teachers felt so.
You also within some education systems, you find that teachers in schools with high shares of socio-economically disadvantaged students or high shares of students with special needs feel a greater need in continuing professional development in the area of integrating digital technologies into learning. So now more recently in PISA 2022, we asked the OECD asked school leaders whether they feel their teachers are prepared to to integrate digital technology into learning and nine out of ten school leaders felt that their teachers are prepared to do so. Now this is from 2022 so I imagine the COVID-19 pandemic has had a significant positive so to speak effect on this.
So you could argue that there is some base for teachers to integrate digital technologies more broadly into learning but AI technologies I would argue are in some ways fundamentally different. In a way, the users understand much less how the output is created based on the input that they provide. Thank you for your thoughts there, Samo. Questions come in from Lyleen Murphy from Canberra who asks, what age do you recommend students?
use AI for their learning? Given we were talking about different kinds of students there in Samo's answer, Punya Mishra, maybe you could answer that one. What age do you recommend students use AI for their learning or which teachers should permit students to use AI for their learning?
Here's the fundamental challenge in that issue is that AI is already embedded in our lives. So if you're using Google Maps, if you're using Google Photos, if you're using Siri, Alexa, whatever, it has already become part of this generation of kids who are growing up. They have AI all around them.
I am not comfortable chatting with my Alexa bot. I never feel like asking it a question. I find that whole interaction of voice interaction to be honestly quite weird. I'm not used to that.
That's not what I grew up with. But then I look at my colleagues and their younger children. They are engaging in conversations with these bots. What does it mean for developmentally for us to be engaging with this?
I think that that cat is out of the bag. You know, we are engaged, our children are engaged with these tools and technologies. So to me, a more important question is really, how do we help them develop understandings and practices that they can recognize, as you know, Madhumalati said at the beginning, what is AI generated and what is not?
What does that even mean once we start getting these meshing of things, technologies and so on? So I think that the issue of whether at what age should we introduce it? That's, you know, in schools. Sure, that's a discussion we can have, but we cannot ignore the fact that these tools and technologies are already embedded in almost every aspect of our lives and our children are interacting with them from a very, very early age. So that's, I think, one thing.
I think I would just push back a little bit to something that, you know, Madhubati said at the beginning, at some point, which is around that this is just a technology or a tool that can be used one way or the other. I think sometimes by taking these to be mere tools. we forget that they also influence how we think and act, right? Yes, a knife can be used to make vegetables and also could be used to hurt people.
But a knife is fundamentally different from a gun, which is fundamentally different from a hammer. You know what I mean? And so I think understanding that AI technologies, by being this generative social, are fundamentally different from previous technologies that have come in the past, let's say the internet or television or radio or print.
And so... I think it's really important as educators for us to be also thinking of the longer arc. You know, if you think about social media, when it came out in 2007, 2008, who would have... I couldn't imagine the sort of positive and also the incredibly negative destructive effect it has had on teen mental health, you know, polarizing conversation across the globe. What this technology does, I think one of the conversations we need to be having is going beyond saying that, oh, we can use it for good and bad.
But there are certain propensities in this technology, which I think are important to recognize and understand. And that's the kind of role that I think we can do here is to broaden that conversation. to say what happens when a technology like this is unleashed on the world, which allows you to create absolutely genuine looking fakes, whether it's video or images and so on.
What does that do to the broader? I mean, one of the problems with AI, for instance, is what we call the hotness problem. Anytime you ask it to create an image of a person, it's always a perfect looking person.
We know already teens suffer from body image issues. What does it mean when this technology is rampant in the world around us? I think those are the kinds of questions we need to be thinking about.
beyond sort of the classroom use, which I think is critical, I get it. But there is a broader story here that we need to be engaging with. Thanks for that, Punya.
I'm glad to say that AI couldn't improve on any of the panelists appearing on the OECD webinar today. So good for us. There is a question here on a very specific topic here about special education needs from Marija Kafkler, a professor from Libyana University.
And she writes, how can... AI help a teacher differentiate and individualize lessons for students with special education needs? I wonder if Mahmati Sharma could answer that one.
So I want to start by saying that children with special education needs are possibly the ones who are also, in my experience, I have not collected enough data about it, but in my experience, are gifted and talented in being able to code better, come up with creative ideas, even though they might get frustrated. So they're actually a good target group to focus on becoming better at coding or AI. In terms of what could be needed, so there are actually assessment tools which would help teachers determine what kind of learning will apply to which kind of student, not specifically for special education needs students, but for students in general.
Some students are better at visual learning. Some students are better at text-based learning. Some students learn by watching. Some students learn by hearing. So that way, the needs of the students could be addressed by teachers using those tools.
In terms of what special education needs students would need, one, there are a lot of tools available for being able to convert. For example, if they are not as fast as writing to convert their text. that they speak out into written transcript form which is needed.
So I know that in Luxembourg and around Europe there are several schools which actually help teachers by giving additional aid to students like this who have a need. Thank you so much for that. A quick follow-up actually because you mentor children and especially girls for hackathons and things like that, I think I'm right in saying for coding competitions.
Has AI changed the way you go about supporting students in this area? Yes, I think AI has actually aided it in a lot of ways. It is easy to say that, you know, even if the students do not know how to code, there are a lot of AI tools which will be able to generate code for them given the correct prompts.
So teaching them about it has made it a little bit easier. The creativity aspect continues to stay, but I think AI is aiding in a lot of different ways. Just as an example, once they have created their... A prototype.
It is possible to get a presentation generated which they could quickly give. As you know, in a hackathon, the time is limited and there are multiple deliverables. You have to deliver your code, you have to do your pitch, you have to deliver a presentation. So AI tools have actually helped to be able to get all of those deliverables out in a hackathon. And many of them are available there.
So for those who don't know how to code as well, there are tools which help in generating the code and making it happen. Thank you for that. Samo Vasek from the OECD, let's bring you back in at a slightly different element to this. Are there any countries or regions of the world which are doing particularly well where schools are using AI more effectively than others are for teaching and learning? Well, there are several examples which will be brought in the working paper, which will be published in the coming days.
Now I could mention a few, for instance, Korea is developing a whole new textbook system where the textbooks will be fully digitalized, they will be instantly translated to several languages, there will be captions provided and at the same time teachers are getting training into how to effectively use these textbooks. In Czechia, in another example, standardized assessments or some standardized assessments are are graded with the help of AI. Now, not solely based on AI, but it's graded with the help of AI and subsequently further checked by teachers.
So I cannot really say anything about the efficiency of use, and particularly I cannot compare different countries in this efficiency, of course, also because countries use AI tools differently. But there are several examples of countries... which have already implemented the use of AI at a more systematic level. Forgive me, I did not mute myself.
Thank you so much for that, Samo. Questions have just come in from Macarena de Castillo, which I think follows on from what you were just saying, asking how can we design effective professional development programs on AI for teachers? Given the vast array of AI platforms available today, what should they be focusing on?
Punyat Mishra, maybe I hand that question over to you. Sure. In fact, I was just replying to her on the chat, so very appropriate. So one of the things that, you know, the work that I've been doing over the past three decades has been around empowering teachers.
And I think given a technology which is such a Rosark block in some sense that we can find a variety of different uses for it. recognizing teacher agency in that space becomes incredibly important. Trusting the fact that they know the context within which their classroom operates, their school operates, and then providing them opportunities to quote unquote, what I call play with the technology.
In fact, we have this idea of what we call deep play. That's actually the name of my research group is a deep play research group, because it argues that what you're really doing is you are engaged in this dialogue, in this transaction, in this conversation. with this technology, with the content that you want to cover. Because thinking about AI in the context of teaching coding, as Madhumalti talks about, versus teaching physics, versus teaching history, there are completely different ways of thinking about it. And the person who's best suited to think and ask those questions is actually the educator at the front line there, right?
And so giving them this opportunity, safe space to play and explore, and then think about ways of experimenting and playing with it. I think we find it incredibly empowering rather than it be a top down saying you have to do X or Y, which A, I think loses out on the creativity. But it's also incredibly disappointing for teachers like that you are not trusting them with how to use this technology rightly.
So I think giving them this scaffolded support to play, experiment and try out what this technology can do is what we have found over time, irrespective of the technology. to be the best way of supporting teachers. Thank you for that, Punya.
And thanks to everyone who is getting involved in the chat. It is very fast moving indeed, and I'm struggling to keep up. And apologies in advance, I'm not going to be able to read out all of your questions. There is one here though, from Anita van der Spool from the Gordon School, who asks, how can AI be integrated in the classroom if the school doesn't have money to invest in IT and pupils are from deprived backgrounds? Mamahmati Sharma, would you like to take on that one?
Sure, I can. So if you're saying that there isn't technology, I'm assuming there is no access to computers, for example, or the internet. There are a lot of tools, unplugged activities, which can be used to teach students about AI.
Just as an example, there is the Turing test where you do an activity with the students. Some of them are asking a predefined set of questions. and people are answering and based on that you have to decide which one is actually a computer and which one is actually a human being being able to answer the question.
So there are actually a plethora of a lot of activities which are unplugged which do not need any screen to be able to teach students about AI and use it. So it's not something which is only for the privileged it is something that can be you know learned by other people as well. And going back to this question for me Someone was talking about the 21st century skills and I would like to bring out that for me coding is a great way to teach the learning and innovation skills which have been defined by p21 the on the 21st century skills so creativity collaboration communication and critical thinking.
All four can be easily taught by coding and incorporated across the subjects. So when I talk about coding, I'm not talking about learning to code or learning a specific language. I'm saying coding can be a tool or an aid to learn your physics, to learn your maths, to learn your English as well. It can be incorporated within the curriculum.
And there are several case studies which we have done country-wise in the AI report that I have linked in the chat where it talks about European countries. and how they have been able to incorporate AI within the curriculum. So for any country who's thinking about doing it, please go ahead and take a look at it.
There are examples from Spain, Ireland, Ukraine, all European countries. Samo's publication will talk about other countries more for, if I understand correctly, learning assessments, right? This one is about how they've been able to incorporate it within the curriculum for students to use. And again, it talks about the different age groups and how it has been introduced.
So I would encourage you to go take a look at the link. I have posted it on the chat as well. Thank you again for that.
Samir Vasek, let's bring you back in, because there's lots of discussions here about what should happen and how teachers can use AI more effectively. But what should the policymakers be doing to help those ends? Are policymakers developing, for example, policies which are going to help teachers use AI or not? What's the picture globally?
Yeah, certainly. Many education systems have already developed guidelines which can help teachers navigate the vast spectrum of AI tools that are out there and also help them to integrate these tools into their pedagogies and into their teaching. I should mention that these are often concentrated in English-speaking countries, even though we suspect that teachers in non-English-speaking countries also use AI tools. and therefore it would be important that other education systems kind of catch up with these. Now one thing that has been mentioned time and again, I'll mention it one more time, is that I think policymakers at the central level need to start pondering the question on how to provide continuing professional learning to teachers and how to integrate the knowledge of when and which AI tools are most appropriate in which situations into teacher training either at the beginning in terms of initial teacher education but also in terms of continuing professional learning.
So I think this is the role that or this is the pressing currently pressing question that policy makers need to start getting their heads around. What is my role from the central level in terms of shaping teacher training in this area? Thanks for that. On the subject of teacher training, if I can come back to Punya Mishra, what's your view on this?
Are schools and teachers being left largely to work it out for themselves? What should teachers be doing if they are keen to learn about AI? Are there certain resources which you would recommend they look at? So we recently did a focus group with a whole series of school leaders.
And the title of our paper says that all, which is the title of the paper, is can we all slow it a bit down? I think that what's happening is that the rapid pace at which this technology is coming at us, that we think it's one thing that tomorrow a new capability or tool emerges in that space, I think is incredibly challenging for teachers, for educators, for school leaders, everybody. So for policymakers, I think that's a fundamental challenge because you don't want to create a policy. which is restrictive in some way, not aware of what is around the corner, so to speak, right?
So thinking of more open-ended kinds of ways of thinking about policy become incredibly important. And I think that from the point of view of educators and teachers, being sensitive to some of the issues with this technology, understanding the nature of this technology becomes incredibly important. But at the same time, being willing to play with it, I think is also as important. So finding that balance is going to be tricky. But...
I think that's really the only way forward. The advantage here is the following, which is since this technology is so new, nobody really truly understands it, right? What its implications are, what its ramifications are.
There are lots of resources out there on the world right now that you can look up, many of whom which have been shared here in this, you know, Madhulati shared a whole bunch of them in the chat and otherwise in what she said. So those resources are there, but it's also going to be, I think at the end of the day, depending on our agency, our willingness to jump in and play with it in a world which is changing pretty fast. So if you think about like, what should our children be doing? I think the same question comes up with teachers that we are all sort of in that same space that is sort of flattened.
sort of the space for innovation in some way, which is to me is a positive because it allows us to ask questions of our practices that we may not have done before. Thank you, Punya. Mahatma, I wonder how much of that tally with what you're seeing in your part of the world? Do you agree?
Is the question about how teachers are using AI in different parts of the world? And in terms of support from policymakers to ensure they have the training? Yeah, so this is actually really good and it brings me to a point and I have been seeing this conversation happening on the chat that, you know, can there be workshops for teachers organized? I think Margaret mentioned this on the chat as well. So, yes, this is an initiative that is actually getting done, I know, at the European level and it is also being done at specific country level as well.
Just as an example, the AI report that I talked about and I have pasted in the chat on the request of everyone. We actually went and conducted workshops on this report specifically for teachers. It consists of seven different reports.
So the workshop was always aimed at what would work for teachers of that particular country. So we did it in Ljubljana, in Slovenia, in Luxembourg, in Germany and many other countries. And we always focused on the report, which would be of most value for that particular group of people. So in some countries, they already have a good level of digital competence.
competencies and the teachers would like to start by teaching their students how can i code with ai how can i develop apps with ai how can actually use ai to make tools so then we actually went in to do a workshop gave them examples of tools they can use like teachable machine machine learning for kids and there are a plethora of their out there as well and For other countries where their digital competency was not so high, the idea was to make them aware that this is the playing field right now. This is what exists in AI. If you want to get started, this is where you can go. So therefore, in the report, there is actually a flowchart which starts with, I'm a teacher who knows, do you know enough about digital competencies?
If yes, follow this path. Do you know about coding already and you want to bring that into the classroom? Then follow this path. So basically, there's a flowchart.
of the different options based on the teacher's competency level. And just to mention what kind of competencies are needed for the teacher is also covered depending on how they want to use AI. Is it with AI, for AI, or among, I forget the last, about AI, what they want to actually teach their students about.
So that could be the way for teachers to actually go about doing it. Okay. Thank you for that.
We are running a little low on time, but there's a great question here from Shweta Singh, who writes, has an AI framework been developed for vocational education? If I can get my words out. Yes. Has any AI framework been developed for vocational education?
Sami Vasek, I don't know if you know the answer to this or anyone else on the panel, but I'll hand over to you. Thank you for that. I am actually not aware of frameworks which would be particularly tailored for vocational education and training. Now, this might also stem from the...
Keep in mind that education systems differ in how they operationalize vocational education and training. So even if they were developed in one country or in one education system, it might not necessarily be applicable to another. But unfortunately, I am not aware of any. I'll hand over to other panelists if they know of any frameworks for that. So I don't have a knowledge of a particular framework for vocational education, but one of the foundational frameworks for technology integration is the TPAC framework, something that is associated with work that I did many decades ago.
And what we have done is now expanded that framework to apply to generative AI. So I would say if you are interested in vocational ed or any kind of specific disciplinary knowledge, then I would recommend looking up, I'm going to drop a link here to a paper that we recently published. wrote around the kinds of knowledge that teachers need to have to integrate AI in their education. And so that's a framework for thinking about how to do it.
And I think vocational ed would fit in very well there as well. But it will require some effort on the part of sort of translating that into your particular domain, but should not be too much of a stretch. And I'm going to drop that link in just a second here. Thank you so much for that.
Purnia, well, listen, we haven't got that much time left. So I'm going to ask you all a final question. It's really to try and help all those watching to answer the question of the webinar. You know, how can teachers use AI more effectively?
What simple steps should teachers consider today to immediately improve? their use of AI in the classroom. Mahanmati, why don't we cross over to you first?
Sure. So I would say, assess your competencies. And what do you want to do with AI as a teacher? Do you want to teach in general about AI for awareness? Do you want to teach the students how to build with AI, which would be the developer perspective?
Or do you want to work with AI to make your life simple by using it for assessments or provide personalized learning and all of those? So depending on that answer, again look up the report which I have linked. It will give you specific resources to use.
I can list 10 different ones down here. But if you go into the report, you will see based on the age group for which you are trying to bring AI into the classroom, there are several resources and it also lists down. We have tried to put in a lot of free resources because we talked about not having enough infrastructure or costs there. And also the ones which are paid. So it makes your life as a teacher very much easier if you follow those guidelines.
But what I would like to end by saying is definitely give it a shot. AI is here to stay. It is, we are surrounded by it all over. So definitely give it a shot. Learn as much as you can about it and keep abreast with it.
All the best on your AI journey. Thanks for that. Sammo, what steps should teachers take to immediately improve their use of AI in the classroom? I think it's...
for teachers and for users of AI more generally, I think it's absolutely crucial to embrace the opportunities that these tools bring with them. But at the same time, we need to keep in mind all the challenges that they come with too. And during this panel, we have mentioned several bias, privacy, et cetera. Now, if you are not quite sure how to do this, seek advice. And just in the chat in this webinar, you can find many, many guidelines on how to achieve this objective.
Now, I want to end up with just one thing which we haven't quite covered. And this is more a message for policymakers, but also researchers and academics out there. We need more research.
In particular, we need more research on what is the impact of artificial intelligence on equity and inclusion in education. What is the impact on non-academic skills? What is the impact on inclusion in particular?
I mean, are we really helping students to be included in the classroom or are we simply, you know, giving them a tablet to be to have a new, you know, intelligent yet still artificial friend? So we need more research. Thanks for that, Samo, loud and clear. And finally, the last word goes to Punja.
What steps should teachers consider today? about using AI more effectively? So I think there's a very, from my point of view, there's a very simple thing. I think there's an incredible push and hype around this technology.
And I think sometimes it's just okay to take a breath and step back. The world is changing. We get it. But I think teachers need to acknowledge and recognize their expertise, their sense of agency. And I think policymakers need to factor that in.
Too many policies are restrictive and say you can do X or not. so on. And I think that recognizing teacher agency, teacher creativity, and that you can go and play with this technology, follow the resources that are out there, and through that, come up with really powerful, productive ways of engaging with what the primary function of school is, which is preparing your kids for the future.
I think to me, that's something that gets left in the way behind because we have all this hype coming out of Silicon Valley or whatever. Oh, if you don't do AI, you're going to fall behind. It's okay.
The world is moving. I get it. But I think giving yourself that sense of agency and power to make those right choices, I think, is the most critical thing. Thank you so much for that final thought, Punya.
And thank you for everyone for watching this webinar. On the panel today, you've been watching Punya Mishra from the Learning Engineering Institute at Arizona State University, Mahamati Sharma from the nonprofit workshop from me, and of course, Samo Vasek, OECD's very own. from the Education for Inclusive Societies project.
Thank you all so much for watching. If you want to learn more about all these topics, there's obviously loads of stuff online. The new OC report that Samo's been involved in, working hard and long hours to the end of the night, every night for many, many months. That's coming out next week.
It's called The Potential Impact of Artificial Intelligence on Equity and Inclusion in Education. So there's some bedtime reading for you. You'll be able to find that. on the OSD website. Thanks again to all our panellists.
It really has been a lot of fun. Thank you to all those who've been working behind the scenes on the production. And I hope to see you all again soon for another OSD webinar. All the best.
Thank you.