What EdTech Gets Wrong About Implementing AI

Key Takeaways

Hi everyone. This is Shawn Strathy and welcome back to another exciting edition of the Tech and EdTech podcast where we discuss technology that powers education and improves learning for all. In today's episode, we're going to be exploring something that goes beyond tools and platforms and gets to a more fundamental question. What does AI mean for education, for teachers, for learners, and for knowledge itself? Our guest today, Stephen Jewel, works across education systems globally and brings a perspective that is as philosophical as it is practical. Welcome to the show, Stephen. Hey, thanks, Shawn. Nice to be here. So, uh Stephen, before we get going, you're currently serving as the global head of AI and EdTech for Teach for All. Uh but I'm a little cur- curious about your backstory. It's my understanding you've worn many hats, uh academic, founder, CEO, advisor, VC. Uh but you have a foundation in the classroom in northern Canada. Can you give us a little glimpse into your professional journey? Yeah. Thanks. That was like I think the the the fastest and probably most accurate representation of like my career I think I've ever heard, Shawn. So, that's brilliant. Uh when you go into this profession, you don't go into the profession with a uh sort of a complicated roadmap ahead of you that that might include any of those nouns that you just set out there. Um but yeah, it's um it began in earnest. You're like like many of us, you go into teaching uh with a view to wanting to contribute to our communities, uh feel like there's something that we can we can bring to the conversation uh and without being too earnest, uh to change the world, which uh to do something that that can make a difference. Uh so, I I looked at a number of different ways of thinking about pursuing that that vision. So, that that that uh that idea that I was hoping to pursue with my with my career, my life, and education seemed like the place that um would be the right right place to to do that. There's a bit of a journey that that led me to education and maybe I'll touch on that briefly here and and then I'll I'll stop and we can we can talk a little bit more about some of those other moving parts, but I think it's it's worth mentioning like there's is a point in in your journey to decide to become an educator where you have to take a view as to whether you pursue that that other domain that you've been contemplating all your life or what have you or that say your your parents are considering or or your your friends are pursuing. So, it's a complex landscape. Uh and for me, I ended up taking a summer job as as we do when we're university etc. And I I ended up working um uh in a job which I hadn't planned on doing that was working in the wilderness in northern Ontario uh working with uh young offenders. Uh that was a was a specialist program run by the then uh um Ministry of Social Services in Ontario and the premise behind it was very much like the Outward Bound model uh that's very well known in in the US. Uh and that is like giving people in this case young kids from Toronto an opportunity to explore alternative stories by having experiences that they may not have had otherwise. And that happened to be in the wilderness of northern Ontario. Uh and it was a real eye-opener for me, Shawn, in so far as like both did I get the opportunity to spend my summer canoeing, white water rafting, climbing, all of those things that were a big part of my out of school life and social life. But I also got to do that with people that were exploring something completely new and different to them. And I think it was for me watching that change that overcame every single one of these these young people uh where suddenly the world became uh a different place, a place that was worthy of exploring, uh that challenged them, uh that presented them with all kinds of barriers to to completing uh the these challenges that were ahead of them. Uh and and that really was really the inspiration behind why I thought education could be a place where the most dramatic change can happen in people's lives, and it was just fascinating to watch that happen. Kind of which then eventually took me to the northern wilds of of the Yukon territory later on after I got my teacher qualification. That's amazing. Coming back to today though, can you tell us a little bit about your role at Teach for All um and what you're seeing across education systems when it comes to AI? That's a pretty loaded question. Uh maybe start with the first half on uh your role at Teach for All. Yeah, so um again, this long circuitous journey to get to this point to work with uh Wendy Kopp who founded Teach for America over 30 years ago and then Teach for All about 18 years ago. Uh and Teach for All is one of those incredible uh organizations that has stood the test of time around its ideas and its vision and values. So, it's it's a it's a real rare thing in the world to to to meet a founder like like Wendy Wendy Kopp that has uh conceived of an idea, implemented it, and it's weathered many headwinds and and managed to iterate and transform itself so that it could deliver across what is now 63 uh network partners across 60 plus countries uh uh across the world. So, it's a really unique uh organization, a global organization of educators with a singular vision about creating an opportunity for some who may not have otherwise come into the field of education very much like the the story that I was telling you about how I sort of came into education, and it creates an opportunity for those who might not otherwise have considered this as a career opportunity to join a movement in which they place themselves in proximity to some of the most challenging questions and problems facing their communities and society and do that work as an educator for for a fellowship or a period of 2 years. I mean, after that time, once completing the fellowship, those educators, those teachers can go on and pursue other things, of course, as lots of educators do. But, interestingly, most of the fellows that graduate and become alumni stay in either social impact and education going forward. So, it's For context, it's a it's a really unique organization where you've got 60-plus countries with educators moving through a program, all sharing the same vision and values about putting themselves in close proximity to the issues that matter in their communities and trying their best to contribute. So, that's really the context. I joined just almost a year ago, so June last year. And that was with a view to working with Wendy around bringing what is the most profound change in education since education was founded, like really. So, there there really hasn't been anything that's comparable in terms of a sea change to the this whole sector and the whole knowledge economy that we're operating in. And so, this opportunity to contribute to the transformation of this organization or really like the like the bringing together facing the challenge of AI through this global organization was an opportunity that I just feel incredibly privileged to be a part of at this point in my career. Um that is fantastic. I think that I know plenty of alumni from your organization that have gone on to work in other uh educational companies, became local teachers here in New York and New Jersey. And I'm always amazed at the perspective that they bring Uh um uh an upbeat can-do attitude to every challenge that uh comes that they that they come across um after they do their time with your organization. So, I think that that's a true testament to uh what you guys do with the teachers uh let alone what you do with the students. Um and coming back to uh you know, you mentioned that this is one of the most This is the most profound change in education uh since the dawn of education which uh that's a that's a big statement. What are you seeing across different education systems when it comes to AI? What's the what's the high-level view globally that you're seeing? Yeah, that's a really great question. Um and this is something that's front of mind for everyone. Yeah, here we are in spring 2026. Um 12 months ago, the the conversation was completely different and 12 months before that, it's was also just very nascent. But certainly today, April 2026, there's not a single educator in the world anywhere at any level who's not thinking about this technology, either using it on a daily basis or contemplating how they're going to either mitigate the risk and maximize the opportunity uh both for their own teaching and learning, but also how their students might be using this technology. So, there's there's a lot there we could dig into, Shawn, uh but the general landscape is we're all in it together. It's one of those really unique moments too where on a global basis, there's something happening in a system change in education in which every single country, every single teacher is paying attention to the same questions. And uh to be to be honest at this point, there is no single right answer yet. I I think there's there's some interesting ways of approaching that opportunity, that problem, uh however you want to frame that. Uh uh that I think is probably the way I would frame it, Shaan, is that we are all thinking about this problem and and that in itself is unique and probably has never happened before, either. I think the speed of change as well as what's different about this revolution compared to previous ones. I come back to like analogies of previous forms of transformation, right? Everybody always goes back to the printing press. I wasn't around during the printing press invention, but the the dawn of the internet and the digitization of educational systems and processes, that that took years and years to to refine and implement and there were early adopters and late adopters and to the common bell curve that a lot of edtech companies used 15 years ago to talk about where adoption sat of of their digital systems. But, this is all happening within like months. Like GenAI really hit the public lexicon just a few years ago and it's impacting everybody around the globe and that just it gives me pause and makes me agree with what you're saying about how this is just totally different than anything in the past. Yeah, no, it that's 100% true and it's I think 12 months ago I think it was not easier to be bullish about the opportunity and and quite excited about it. You know, who wouldn't be excited about open access to knowledge for every single person in the world wherever they happen to be with the opportunity to innovate and create from that knowledge history. That that's kind of a remarkable thing, right? And and I'm I'm still very very excited about that. But, I think, you know, what you and I are both alluding to here is like the problem is like it's not that this I mean, this is this general purpose technology which is probably, you know, the deal the single other greatest, you know, technology on par with this is probably fire. You know, like the In fact, recently here showing I should say UK, they were doing a dig an archaeological dig where I think they were discovering or claiming to have discovered the first location of like human-created fire, you know, at least you know, on the on the side of the ocean. I think that's fascinating. I I think you've touched on the thing too that that is the most troubling and I know in this this our conversation we'll we'll get past the big concerns and maybe on to some of the the big potentials. But speed is a problem um for us, that's for sure because humans generally are are cautious by just by by by history, by our experience. And that's particularly true when it comes to young people, parents, children, teachers, and schools. Uh they're rightly cautious about change. And so that that's a real real dilemma. Um if we were talking about banking, I think we'd probably be having a different conversation. Um same same dilemmas, but our way of thinking about the existential issues would be framed differently. Uh but when it comes to humans and particularly young people, um caution requires a pace that is perhaps not at the pace of change that we're seeing right now. So that that presents a serious problem for us, I think, as as educators. You just used the term a moment ago, general purpose technology, um describing AI. Uh and it and you've said, I think, in the past that it takes the shape of the person using it. What does that mean practically for education systems, teachers, students? You can take that in any one of those three directions, really. Um but what do you think it means practically for for us? Oh yeah, you see this thanks for um mentioning that. Um it's it's a great question and something I'm I'm thinking about and uh the great thing about this technology is that since we're all thinking about it, every time I encounter someone who's also discussing or exploring these these questions as as are you, uh that really all we're doing is all level setting as where where we are on that journey and what are the the front of mind topics. Uh but for me right now is this idea that um as a general purpose technology, you know, it really does just take the shape of the person that picks it up. Uh I I think that's both works in metaphor uh as well as in terms of real practicalities. And it's also presents a lesson. Um the the practicality part is that um you know, it enables me sitting here in Cambridge, UK to work with the tools in a way that are meaningful in the context of let's say my household. I'm not a practicing educator at schools right now, but I am with my own kids. So, I'm able to like use those foundation models, ChatGPT, Claude, Gemini, whatever your flavor, uh in a way that's contextually relevant which you my specific kids learning needs at my house. And so, that I think is really powerful, you know, and and so, sitting somewhere else in the world in a different context, I could use the tools to do something different in a different context, in a different pace, in a different way. And I think that's incredibly powerful and important. Um and this leads to another part of like this idea of these the foundation models as being contextually relevant to the to the person that uses them. There's two things that I think about in that framing uh makes that's particularly important for educators and teachers. Is that in order to use these tools effectively, you you really have to do something that educators, master teachers, inherently understand and that is you have to know who you are. And and by that I I don't just mean euphemistically, just recognize it simplistically simplistically kind of who you are in the world. But you really have to understand what your motivators are, what you're trying to achieve. Uh I'm see speaking about this in terms of like as an educator, as someone who's thinking about like the teaching learning process and pedagogy. You have to understand who you are in the landscape of what you're trying to achieve as an educator. And also within the context of the content knowledge that you're trying to explore. So, whether it's math or geography or learning new language, whatever it happens to be, or even just exploring the big big questions. Without having those foundational elements in place as a person is using this tool, it that's where the risks really start to emerge because it doesn't allow you to make um informed decisions or informed critical analyses about how you use this incredibly powerful tool. So, there's there's a couple of things about that. You know, like yes, this tool reflects you know, each individual use, you let a hundred flowers bloom kind of metaphor. It's different use cases, whoever picks it up. But also the the the effectiveness of how that tool gets used is deeply contingent upon that person's I'll just maybe I'll use the word readiness, you know, to use that technology. And the readiness comes back to that that first part I was talking about around understanding, you know, who you are in the world in relation to the what you're trying to achieve and learn, and also that you have, you know, a canon of knowledge or at least a foundation to the canon of knowledge that you're trying to explore to allow you to make discerning questions or ask discerning questions of this technology. Otherwise, it will has the potential to to to to to learn you, you know, as opposed to the other way around. And we we we don't want that to happen. Yeah. Yeah, I mean, just on a general personal experience level, um I you know, cuz obviously I use chat GPT, I use Claude, I use um you know, various tools whether I'm doing it for my professional work or my personal life. And I sometimes when I try to switch which model I'm talking to, um it doesn't give me the perspective, the feedback, or or anything that I'm really looking for. And this is kind of funny. I use one model for uh all the coaching that I do for baseball here in New Jersey. And uh it has a very congenial tone. When I try to do work within that model, professional stuff, it's spitting back stuff at me like, "Hey, bro." And I'm like, "No, no, no. This is You're talking to the wrong the wrong version of myself." And I'm I'm not identifying with it where I am exactly, I think is what you're trying to say, too. And there's a level of of of consciousness on how to use the tool and not let the tool adapt to you. Did I Did I kind of get that? Yeah, you know, that's really interesting, Sean. You're just kind of like expanding what I kind of thought here because yeah, you were at we are actually we apply these tools. We are kind of not just a single person in the world. That's what you're kind of There's not There's not the There's more depth to to us each of us than one one moment in time and how we apply these tools. That's that's actually really brilliant kind of addition. I love that um that Yeah, because you you're as a as a baseball coach, you definitely like working a different vibe than when you're doing English lit or what have you. You know, or even like the expanding the getting up on stage and about to present a keynote on a specific topic or what have you. There's There's different voices, isn't there? And that's that's actually quite fascinating. I like that idea a lot. Yeah. Yeah. Um and so uh coming um to our next question, you work with a wide swath of of people across across the world. And um what is the biggest mistake you see education systems making when they're implementing AI? Yeah, okay. Um so for context, too, like we and So there I'll I'll get to the I have a view on on what could be some of the biggest potential mistakes with this tool. And I've sort of touched on them a little bit already with some of the questions that we've we've covered here, but um I should say that it's still so early in terms of how schools and and educators and parents are acclimatizing to this this new paradigm. Um that we're in that that that point where um the mistakes are small. Um so we we we can see this in the way people are talking about the risks to to learning um the risks to authenticity in terms of like homework being submitted etc. from students. There's there's all there's these what feel like I think right now like let's just say that two three years from now Sean when we look back at this this this time spring 2026 we're going to think woah those those those questions that you were asking or struggling with at the moment Sean those weren't the real problems you know like we're we're we're well past that we're I think this might seem I think we're moving so fast that these these problems are going to seem naive at the time. But I think I think it's a cutting right to the to the point here. I think the biggest mistake would be that there is one way to do to deliver this. Okay. And I I think the the other like and let's not talk uh frame it in terms of a mistake but I think the way of like mitigating against the risk of going in a direction that may not lead to an optimal um inclusive equitable um accountable safe implementation of AI is to is to to do a top-down. Like this is what we're going to do. This is the model we're going to use. Here's the EdTech that we're going to use that uses AI. Now let's all just do that. I I think if we do this if if we do it that way Sean I think we're just doing old EdTech. You know and I think this technology is so incredibly new and so untested that to implement an AI strategy the way that we used to implement rollouts of new EdTech pre-AI would be a massive mistake. So maybe that's the way you've actually forced me into like a way of conceptualizing this. Let's not replicate let's not use that same procurement deployment professional development cycle that we used pre-AI. I think that would be a massive mistake. So what we're thinking about at Teach For All across our network partners is we have this incredible advantage of having a diversity of ways of thinking and knowing, of education systems, of curricula, pedagogies. You know, the list is long and how we think about teaching and learning across this this global network is that we're out there like on a big exploration and listening tour with our with our network. And the thing that we're hearing and what we're seeing is that teachers are interested in exploring all the models, unwrapped models, models that are the foundation models again, the Clouds, the ChatGPTs, the Geminis, etc. Some of the the edge models that are coming out right now are being explored too by some of our network. And those very conversations that they're having and and how they're exploring them are getting shared across our network. And that is informing this next phase of potential, if I had to like if it had to be defined and it will be at some point, some implementation strategy that's coherent, that's sustainable, that allows us to do this effectively. So, I think I sort of touched on the biggest potential mistake and the biggest opportunity I think lies in that must just frame it as trusting educators within the paradigm of supported guidance by the systems that run schools to explore these tools responsibly so that we can begin to understand what are the effective implementations, what how can we use this differently in different contexts, by different subjects, by different ages, of course. So, I I think that that that I think is the direction that we'd go, Sean, is like I think we're in the exploration phase right now. I think based on what you're saying, I have two conflicting thoughts in my head. One, I hear you on we don't know what's going to happen in the next 12 months. Like you said, we could be we could have this conversation next year and look back and say our our predicaments that we identified were silly. Like >> [laughter] >> we moved past those so quickly. At the same time is and so you want you want people to continue to explore and and try things out because we don't necessarily know where it could go and it we're only going to find out by people experimenting. At the same time, we've seen general technologies like cell phones, social media, if you wanted to consider those general technologies, but just widely available consumer technologies do a lot of harm on students as well. And I think that there's the concern about AI doing uh causing more harm than good in the educational setting if not rolled out uh strategically and thoughtfully. So, what would your thoughts be around like healthy student use of AI? What does that look like and and can we put guardrails up today that can at least allow for that innovation to happen and that experimentation while keeping kids safe? Yeah, it's a really good framing. Um short answer is and then I'll give us a little bit of kind of you're triggering some thoughts here for me as well too is Um short answer is yes to can we and should we put in guardrails to support student exploration of this technology alongside teacher exploration. Most of what I'm thinking about Sean too, of course, at Teach For All of us we work directly with teachers and and how this gets deployed within classrooms is another layer of that question. Um so yeah, 100% I'm a big fan of of guardrails. So, what that looks like in terms of like but now define guardrail, define who's setting the guardrail, what are the parameters around that? That that gets into that domain of like where the constraints lie and who are the decision-makers around those constraints. And my default would be towards the educator side of things because as an educator myself, as a teacher, I generally trust teachers. I I don't think I've ever met a teacher in in my life now standing the rare few of course that doesn't have the best interest of their of their students in in mind and in heart as well as their parents and their community. So that's a that's a universal across this profession. And so we're really fortunate to have that that that incredible group of people out there in every single country highly motivated by the same core principles. And so I I think teachers are going to do us a a service by allowing them to be deeply involved in this and that comes back to that previous question is like I think this is a time for exploration by educators to explore how these tools might be safely and both pedagogically effectively used in classrooms as well as in terms of some of those safety and health issues that you're alluding to around social media. But before I do that I I I think the the concerns the parallel concern between social media and AI are [clears throat] are are quite separate actually. But let's remember that um social media was never intended for education um and it sort of found its way into the vernacular onto phones and then worked its way down into younger people and younger and younger people and that sort of happened without us really paying attention as as a society. The advantages that we've gone through those hard lessons of social media with young people that were now eyes wide open on big and powerful changes in society. Um I would argue again that Frontier Labs are building a technology that is not like on par with what social media was designed to do. These are two different purposes of these technologies. But perhaps this this period of learning from the impacts of social media and its implications for society and how we communicate with each other where we spend our time uh is as I say to my kids sometimes, a cheap lesson before the real potential issues arise. And so, so maybe maybe that's maybe we can learn some lessons there, Sean, and make sure that we pay close attention to this technology that is really designed for us to understand the future of learning and creativity and innovation going forward where social media was never intended for those purposes, of course. Yeah. I think there's concern, rightly so, by users of the foundational models that these big tech companies will slip into the same business model as the social media companies, gathering information about their users and selling it for advertisers. And so, I think we are more wide eyes wide open now than we were 10 years ago on that type of fact, 15 years ago. Um, but yeah, I think I think it's there's definitely guardrails. I hear you on the teacher trusting the teachers. Couldn't agree with you more. I mean, nobody teaching for fame and fortune. It's everybody who chooses the profession chooses it because they care strongly about their students and the community in which they serve. So, coming to the teachers, again, what do you think are two or three teacher skills that will become more valuable in a post-AI classroom? Um, well, I think one of the great things that teachers have always brought to the table is is creativity. Um, and I sort of just like exploring this idea that if the future belongs to creatives and teachers are naturally creative, then the future belongs to teachers. You know, and that's a little bit didactic, I know. Um, but I think that, you know, as a general purpose technology, when you're trying to problem solve and you're working in the knowledge economy, and also the creative economy, teaching is a creative exercise every single day. So, not only are you dealing with complex content, and that's that's true of early years education all the way through, of course, higher ed. Uh just different way of framing ideas. So, Stephen, we've built education systems around this idea of knowledge scarcity. And I thought about that question hard before saying it, but I still think it holds true. The internet changed it to an extent. AI challenges it way further, and it raises some hard questions. What happens to the value of knowledge when every student has near instant access to explanations, examples, and synthesis of ideas? Yeah, it's great question, Shawn. Um yeah, and and so education's always been in the business of of knowledge, of course. And we think about the knowledge economy and the existential risk to that economy. We actually aren't often in the news and media reports thinking about educators and schools. I mean, that usually goes back to all of those jobs out there that are like living in the knowledge economy, whether that's from lawyers to to to to accountants to administrators, et cetera. We often think about the knowledge economy in that landscape. But in fact, educators are operating in the knowledge economy every single day. And that's really really the core purpose of what schools have always been about. I was talking about this recently with with some colleagues, um and that is like when you strip back the this idea of like if if the if the transaction between or the education Let's call Let's Let's work out a round on this in this regard. If we if we think about schools as holding a education contract with with with with learners. So, you come to school, you've got your educator there, your teacher's there, and they're they're to support you on your learning journey. Now, if the teacher is no longer the gatekeeper of that knowledge, and you can access that knowledge from anywhere, from home, and perhaps even a breadth of knowledge beyond what your what your teacher can provide, what's left? I think that what's left is the part that educators have always prided themselves on over the years. Is that it's not education education and teaching is not necessarily this contractual arrangement around imparting knowledge. We've often seen it that way. Go to school, get knowledge, do your homework, revise for your exams, sit your exams, and so the loop continues. And maybe what we're doing is we're heading heading towards you know, a period where educators will revert to the the pedagogical, the understanding, the nuanced way of understanding how each of those learners can can get accelerated in their own knowledge exploration. And I think that's perhaps the the the next like frontier for education is moving beyond the knowledge economy that education's been involved with for all those years, and more to the pedagogical and learner support engagement that that's naturally inclined towards education and the huge part of what educators do and value every single day. I think you hit upon why I had a hard why I thought so long about the question because um this idea that the agreement is that the holders of knowledge are going to impart the knowledge onto the learner and the learner's going to go on with benefit of having that knowledge now to get a job, whatever. Um that really hasn't been the case for a long time. Uh and maybe we just as educators and people in education need to do a better job of promoting what modern pedagogy is really all about. Um uh there's this uh you you touched upon like almost a Socratic method of teaching, right? Just guiding the guiding the learners on their own learning journey uh for them to um achieve their goals, not necessarily this transactional uh transfer of knowledge. And um obviously that's as as old as time, but I feel like AI um and and and especially not knowing what the future is going to hold, places so much more value on those teachers who are able to help those students learn how to learn, number one, but also even in the absorption of information, make them the the neurological connections almost like working a muscle, so that the next problem that they face, which is going to be different, it's not going to be the same problems we're facing today, they have those synapses firing in a way that they can then solve those problems in the future. And um yeah, I don't know where exactly I was going with that, but just as you were talking, this is what was was ringing in my head. No, it it's a it's really great, actually, cuz again, not being in the classroom every single day, even though I sit on a board school here and and try to stay as close as I possibly can, and I work with educators worldwide, where you're not like right in a classroom and watching that learning happen 6 to 8 hours a day, 100 plus young learners every day, kids coming in and out of your classroom, where you really kind of like see this stuff happen. I mean, the places where I I see it are at the kitchen table here for homework, and my kids struggling with various topics, getting for exams, etc. And you can start to see like little thin edge of the wedge of the change happening around this knowledge economy boundary and what kids value in in when they go to school. And so, here's a kind of a real example. Um of course, in education we encounter many amazing teachers, right? And sometimes we encounter teachers who are either teaching out of subject because a school may lack that specialist or what have you, and there can sometimes be gaps, like I say that generously. There can be there can be gaps in both curriculum and content expertise, and the ability to allow all learners in a classroom to be able to access that content, particularly in some of the those building a STEM-based subjects, where if you miss a a learning block along the way, it can present some real issues. I mean, one of the things I'm seeing in my own kids here at the kitchen table is that their anxiety level has come down significantly around this dilemma of like, I missed something or I feel like I didn't understand what my teacher was talking about today. I don't really hear that as often, actually. I actually Hang on Heart is one of the biggest changes I think I've I've seen from my own kids in their learning is that what they recognize is that it's not a problem for them if they missed some element of both the knowledge transfer or the or the logical understanding or even perhaps even the implementation of those two things in a novel way, because they recognize that at least in our household, they're able to access that guidance either from their parents, of course we're deeply involved, the relational side of this we should talk about this at some point too, is incredibly important. But I would say that um one of the greatest like outcomes from this is that I think I'm seeing, and I wonder if this is universal actually, um uh kids feeling a little less anxiety or concern about missing those moments in class, because they know that they can have it explained in a different way, in a different language, in a different pace, until they find that that moment. And I think that's that's deeply interesting and probably a great relief to educators, I would say as well too. Yeah. I think um uh uh I want to ask you a couple of more questions. Uh I've been holding you for quite a while here. Um but there's a couple of important points that I want to uh get your thoughts on. One, we we're dancing around a little bit of this question where um students are using AI. It's It's a It's a um an inevitable fact, right, that students are going to use AI. It's It's um ubiquitous across our our spectrum here. Um what is the risk of outsourced thinking? So, or where does that lie? Like, how how do we avoid that risk or mitigate that risk of students just outsourcing their thinking to AI Uh during their educational journey and actually making those connections that we were talking about before. I think it comes back to your point re social media and some of the mistakes we made there and are continuing to make. Is that it's such a passive like social media is completely passive. It's really consumer based where you consume things that are completely unrelated and and don't support you on a personal journey really in any meaningful way. Initially like I shouldn't be too hard on social media. There was probably a time in our midst where I think we all felt quite hopeful about being connected to friends and family around the world and and it still provides some of that that service. So moving towards this idea like about sourced thinking like directly and and whether or not that's that's going to be a problem or whether we're seeing that. I think you're you're right that it's something that's front of mind because we don't really know yet. I think both in terms of how students are working in a pedagogical framing in classrooms at home with their homework. I'm not even going to get into the neuroscience of it because that's well outside of my scope here and so but I'm sure there's there's going to be studies that are going to help us understand this. But I think just from an educator side of things um where we have the tools that are moderated by educators again coming back to the the start of our conversation you understand the the benefit and utility how we can use these tools responsibly uh thinking about those core of understanding your applied kind of knowledge how you think about things the hermeneutics or your own logic yeah the ways of thinking of knowing of working with the technology and and ideas. I think by doing that these tools will allow us to like think deeper and think more broadly test our ideas. Uh but that will come through again responsible use of the tools and this comes back to that most important point, which is by having teachers in the center of this conversation working with these tools to mitigate the risk and maximize the opportunity. Educators, I think we're going to put us in a strong position to ensure that we can we can reduce the risk of of the of shifting the responsibility of learning outsourcing as we say into some AI model. This leads into my next question that I well, really two questions. In the near term, how is AI going to impact assessment approaches? But also what does this mean for EdTech and the way that EdTech companies build products today in a in an AI world? Which I think our audience probably is most keen keenly focused on. But you just said something that I don't think I'm seeing in the market today, which is this very much still instructor led or instructor oversight instructor guided use of AI. The way that AI is implemented a lot of times today in in educational technology products is that it's a standalone separate feature. Right? The student is interacting with on their own with an AI model. That AI model might be trained based on the content. It might learn about the student, might help them individualize their instruction. But what it's lacking, I haven't thought about this too much, is that instructor oversight and that teacher oversight. So, I'm sure that blends together my questions too, which wasn't a great way of asking you a question, but what are you thinking about when it comes to AI and assessment and AI in EdTech products? Yeah, thanks for that. Yeah, narrow two really big and separate ideas. Um, I'll start with AI and EdTech. Um, one of the the thesis that we're exploring here I teach for all is this idea that uh AI foundation models all chat GPT, Gemini, etc. um present this opportunity for educators to begin to develop the kinds of tools or implementations uh that they might otherwise have had to have purchased or had built for them somewhere else in an in an edtech product. Now, that could be like let's the simplest not simple not simple in terms of an important but actually simplest in terms of like technically feasible. Let's look at just changing uh the the reading level of a piece of text. Let's say it's a we're in topic of geography and in a in a classic any typical classroom anywhere in the world, not everyone in that class room whether it's 30 or 100 students in that class are reading at exactly the same level and we're requiring the exactly same amount of time and do the exact exact same like amount of time to be able to to do that work. Um right off the right off the shelf, these foundation models can support teachers to explore the possibility of differentiating that lesson and I think that's brilliant. So, the framing here is that we're shifting potentially from a dependency on third-party providers around edtech to create the solutions that educators need today depending on what the needs are identified in their class and their context using these foundation models right off the shelf. And I think that's a really interesting space and it really lines up with how educators think about their roles both as creatives and as responsible pedagogues that they can now differentiate, adapt, create tools and assets that they otherwise might not have been able to do because they didn't have either the technology or the tools to be able to implement this. So, I think that's that's step one is I think there's a real shift from dependency on edtech to an agency of educators to be able to do these adapt adaptive ways of learning with um with these tools that were otherwise required by by a third-party. In terms of like AI and assessment so it's also it's both a similar question but also different because it's the problem here is if we're using AI to support our work, whether that's creating a lesson plan for students to be working from, and that's that's the teacher side of things, whether you're creating utility for students to do a end of unit exam and practice at home, whether it's some sort of artifact you've created as an as an educator, or whether it's a student using AI to support their learning at home while they're advancing, going back to my earlier comment about filling the gaps at home. Like where there's a reduced anxiety, I think, by students who know they can access both the knowledge and the logic of what is that they're learning through these tools. So, what happens when you remove those tools? Like let's say AI disappeared tomorrow. Well, first of all, that's sort of sort of like an idea, but but where it really happens, and you and I both know where this is going, is like at the moment of assessment, right? So, I'm not talking about like homework cuz homework isn't serious. There's a real dilemma for teachers right now at homework. But when it comes to actual assessment, like we're talking about summative assessment moments, I think we're going to see pretty rapidly, and educators are the best positioned to be able to do this, too, on a week-by-week, month-by-month basis, is to be able to understand whether they're seeing different patterns in terms of assessment outcomes of individual students that they know really, really well and have known for years in their school and or across the spectrum of their their classroom and school. So, I think there's a a ton of interesting data and insights that are going to come out and probably are I mean already across schools around the direction of travel of of assessment outcomes, summative assessment outcomes by topic, and that's an area I think we really need to pay a ton of attention to. And at the same time, as we look at those outcomes, we should probably be in parallel, of course, looking at what are the things that those teachers and schools are doing that are providing presumably better outcomes and if the outcomes are are are sort of not you heading in the right direction then also looking at those really cut carefully and closely and making sure that we can we can we can adapt and adjust. And I think that's that's how I would look at this is like really a part of that sort of constant iterative agile environment and I think teachers are up for it and it's just a matter of like schools and schools administrators really like they've got the the the the capacity but and trust I suppose of their boards to be able to go ahead and and and and try these things out. Well, I think we we would all be good educators to focus on outcomes as we started the conversation I think on outcomes and we're kind of coming to the conclusion focus on outcomes as well. So I think that that's a great point to to to wrap start to wrap up. I have two lightning questions for you. Just a little bit of fun here. What's the most overused phrase in AI and education right now? Oh, okay. Uh Actually, I was just talking about this with a friend of mine just last week. This is going to be like semi like controversial. So Sean said there's context to it. But human in the loop I hear human in the loop like all the time, right? So I'm a bit worried. I mean I know what that term means. I 100% back the term but it's certainly overused and anytime a term gets overused we know what happens then. So I think we need to be really careful about making sure we understand what human in the loop actually means. So there's my first on on that one. Hot take. Um no, that's great. And then what's one thing that the future that you feel optimistic about with the future of learning and everything that's changing in the environment with AI right now. What's your most optimistic thought? I just I love this idea that again again notwithstanding barriers to device data digital access. This idea that just about any learner anywhere in the world have access to the same knowledge. I think it's just just truly amazing. Like that that is just unbelievably universally important, I think. And so so long as we can continue to make sure that moves in the right direction and that everyone has access to this to this new way of understanding and having access to global knowledge, I think that's just a brilliant change for the world. Well, that's a great plug to to wrap up on because Magic EdTech's tagline is digital learning for everyone and we absolutely agree with that. Um Stephen, just before we wrap, is there is there any place that our listeners can find your work and stay connected with what you're doing at Teach For All? Uh yeah, drop by teachforall.org anytime and you can get a sense of the amazing people that comprise this network of educators across our global network. I would actually encourage people to reach out. We're always happy to engage with teachers and educators, everyone and who's interested in the educational ecosystem worldwide. So definitely find me there. Sean Yelland there's there's lots of great stuff we can share and amazing colleagues to be introduced to. Well, Stephen, this was a great conversation. Thank you so much for your time today. Thanks, Sean. Great fun. >> [music]