The Case Against Rushing AI in School with Pip Sanderson

I'd been to a UNESCO conference in Paris. It was posited as though AI had changed education. It was a done deal. Thousands of professors and ministers from around the world talking and debating, but all about how to do it, not whether or not we should be doing it. I can't believe that actually was the position that was the line. It needs incredible nuance to unpick it, and there is no, it's definitely bad, it's definitely good, it's hellishly complicated. Welcome to the Expert Intelligence Podcast. Today's guest is Pip Sanderson, who brings a balanced perspective from the front lines of AI and education. While the ed tech world rushes forward, Pip challenges us to consider what's at stake in our pursuit of technological efficiency and how we might preserve the human essence of learning. Pip, welcome to the podcast. Thanks for having me. So before we get started, I just want you to take a minute and introduce the audience to who we are and what we're doing today. So I currently work for the National Institutes of Teaching in England. We were set up a couple of years ago, and my role is to design the training that is given to people that are just beginning to become teachers, people in their early career stages, and those who are more established, thinking about becoming middle leaders, head teachers, executive leaders, and those who are more established, and those who are more established, head teachers, executive leaders, and so on. And the work that I do is about taking the evidence about what we know, about what makes effective learning, and thinking about how we best teach that to teachers, and how do we bridge that gap between what the evidence tells us and how you apply that in practice in a real world classroom. And I'm an English teacher originally, and that was my background for years, and before that I worked in advertising, and I always think there's a huge commonality between those two worlds. So advertising basically was trying to convince people to buy things they didn't want to buy, and being an English teacher, in some of the schools I worked in was convincing kids that didn't want to learn, that it was a good idea to learn. So there was lots of crossover there. But my recent work has been thinking deeply around technology and particularly around AI, and trying to figure out what on earth this means for the way we train teachers and leaders, and for the way we teach our young people. So you're neither anti-AI or pro-AI. You advocate for a more nuanced understanding. Can you share the aha moment you had where you're like, oh, this technology is gonna be impactful, we need to think about it? Sure, so I wish that there'd been like this really good aha moment. It's more akin to wandering through this swamp with swirling mist, and occasionally the mist parts, and you get a little bit of clarity, and then the mist closes around you again. So it's this roller coaster journey of thinking it's really significant, thinking it's really dangerous, thinking it's gonna be brilliant, thinking it's gonna be really dangerous again, and it's that continual journey. And I think for me it's been two and a bit years, and the way I described the journey was one first of fear, and then as I learned more, it became excitement. And I'll just talk a little bit about some of the work we did. Back this time last year, so we ran a couple of generative AI hackathons with schools and leaders, and it was like a training method to help them learn about the technology. To take a current challenge of problem that you're facing as a teacher or leader, work together in a group with a facilitator, using a large language model to try and see if that can help solve your problem. And I was there as the facilitator, like professional developments, my thing, it wasn't particularly the AI that I was there for. And I've delivered thousands of hours of professional development in my life, but I have never seen such an emotional reaction to any form of professional development as I saw in those rooms over that course over the last few weeks. It was open mouth, shock and awe from teachers. And I'll never forget, there was a particular teacher who worked with children with special educational needs who literally had tears in her eyes, and she said the phrase, I can get my life back now. And she was talking about the way chatGTP, that's what we were using, would reduce the administration in her role so significantly she would get her life back. And it moved her to tears. And it's really hard not to get caught up in that potential, that opportunity. It was exciting and thrilling. Also, imagine balancing that massive sense of emotion, the room and sense of potential, with the fact that when I look back at the information we were given around the sort of ethics of AI, we only skimmed the surface of it, but you don't know what you don't know. So the kind of diet of ethics that we were given during the hackathon, like I thought that was the ethical consideration. I could have kind of made peace with it and thought it was fine. It's only like 12 months more of reading and thinking, you realize there was actually much more ethical considerations when you dig a bit deeper. And also, I look back and think when we did that hackathon, there was like a, what felt like a really detailed explanation of how the large language models work. But again, looking back now with 12 months more knowledge and experience, I realized it wasn't. It was posited as being a good detailed explanation, but I don't think it was at all. And I think basically with poor information, a lot of positive emotion in the room, I got caught up in that potential. And I think teachers are typically problem solvers and there's a lot of problems to solve in education. And like a restlessness to always want to fix stuff made me and I think other practitioners kind of over excited about the possibilities. Having said that, there was something always like niggly at the back of my mind about the possible unintended consequences. But I really believe they could be mitigated for. And I think I just so desperately wanted there to be a fix for the problems in education that I thought that we could probably put the kind of guardrails that were needed into place. But 12 months on, about 10 months on, I came across some different writings and different musings of very intellectual, clever people that kind of exposed me to a huge stack of information I didn't know about AI and made me realize the more I knew, the less, the less I actually knew around it and had so much more to go. And I can reference the work of a fantastic thinker and speaker called Ben Riley. And he has a think and do tank called Cognitive Resonance and we'd worked with Ben a lot in the past because he's been very influential in the cognitive science space for many years and his previous information organization is called Deans for Impact. And they've been really helpful in shaping the way we design professional development. And he was the first really dissenting voice that I heard about saying that maybe we don't need to do this whole AI thing in education. It was literally the first time I'd heard someone say we don't have to do this, it's not a done deal. I'd been to a UNESCO conference in Paris, like UNESCO, the whole aim for that organization is like global peace. And at that conference in September, 2024, it was around digital education. It was posited as though AI had changed education. It was a done deal. There was no one had to say in it. It was just gonna happen, let's just all get on with it. And I think looking back now, I can't believe that actually was the position that was the line. And for two days of thousands of professors and ministers from around the world talking and debating, but all about how to do it, not whether or not we should be doing it. So it's been a journey, it's been a roller coaster. There hasn't been one moment, it's been utterly confusing and remains confusing to this very day. And that's why I think it needs incredible nuance to unpick it and there is no, it's definitely bad, it's definitely good, it's hellishly complicated. What are the specific concerns that you're starting to see when people come and say, it's time to implement, let's go, here's the how? What are the things that emerge from research or from your personal experience that aren't the efficiency gains of a teacher is that saves them time because they're using a model? Okay, so there's a couple. So it can save time, but what's the consequence of that time being saved? Do we really think that person is gonna do less work or they're just gonna be asked to do more work? I think we know from looking back in history, when new tech comes along, we think it makes us more efficient and freezes up more, we don't, we just get busier. The expectation of the workload just increases. So I think that's a whole misnomer that we're gonna save people time. We're just gonna move around how we spend our time. So I think that's an issue. But I think the bigger issue for me, when people say, let's implement this into education, it's based on what? Based on what? Based on the current tech right now or based on what everyone's telling us the tech's gonna do in a few years' time. Because if it's based on what the tech can do right now, then we know that's pretty limited. So if a student was ever gonna use a large language model, like we know that could often be wrong, like 20% of the time, that could be wrong in what it tells that young person. Are we uncomfortable with that as educators? And the fact we're probably paying for that subscription for that young person, so we're paying them to be getting wrong information at least 20% of the time. And the worst thing about it is they wouldn't even know it was wrong because the AI doesn't think it's given a wrong answer. Like we call it hallucination. It doesn't think it's hallucination, if it could think. It's giving a perfectly grammatically correct answer in its mind on the data that it's been trained on. It's just not correct in our world. And I think that's really dangerous. And one of the other things that I find really concerning is this sort of artifice that's being created in the way the companies want us to interact with their chatbots. So rather than being like a straightforward transaction, it's being gilded and shaped to feel ultra human. And it unnerves me massively. The way like chatGTP now interacts with me, like it's bordering on flirty. Like it complements the way I ask it questions. And it gives me winky face emojis and encouraging thumbs up. And like, I don't want that artifice. I didn't ask for that. I don't want to have that sort of pretend relationship with the computer and I'm an adult and I can push back and say no. But for a young person, that's much more impressionable perhaps. Do we want them building these synthetic relationships? That is definitely not in the interest of the child. For me, that's in the interest of the companies because they know that's gonna probably make their product a lot more addictive and end up generating increased revenue for them over time. The other thing I would say on this as well, which I think is a massive concern for young people is we know a lot about the way people learn. Like there's been decades of research on cognitive science. We don't know a lot about what impact using large language models or different forms of GNI has on young people. Like there is an emerging evidence base and it's mixed and it's confusing and it's really hard to stay on top of, but there is no real clear consensus. And there won't be for years because impact of learning takes years and years to come to fruition. So we're making this like big bet with all this time and energy that schools are pointing towards figuring out how to use this technology. And we don't know whether it will actually do the most important thing of helping children to learn more, be more confident, comfortable in their own skin, enjoy the world more. Like we don't know any of that stuff. For every possible learning gain there is, and I think there are some, I'm just very worried there are more side effects that we have to try and mitigate for. We'll have to try and overcome in the future. It's interesting to me that we've seen this technological advancement before in social media. As it relates to kids or as it relates to us where there was this promise of we'll all be connected. We'll all be connected and everything's gonna be great. And 10 years down the road, 2015 is when the algorithms were updated at Facebook to be more engaging to your point to drive revenue and growth and all those sort of things. And the unintended consequences of that technology on kids, on us, on democracy has been pretty concerning. It hasn't lived up to the hype of what the PR agencies at Facebook and others were telling us. It feels like there's a pressure cooker in decision-making where the companies are coming in with dollars and solutions and really trying to pressure school districts and educators to adopt this technology as quickly as possible. How are school districts handling that inflection point or that push? I think it's really mixed. And I think those that are trying to embrace it are doing it from the very, very best of intention. Like they genuinely see this as being a solution to some of their problems, but it is difficult to make decisions in that pressure cooker state. And I think the schools that are embracing it are, there's that potential that they're mistaking speed and structure of learning that the AI promises for depth and meaning. And that sort of quick onboarding of the technology is just gonna push this really far down that path of efficient learning. Now, I think efficient learning sounds great. Like who doesn't want their kid to learn loads and do it in like super speed time? That sounds brilliant. But we know like that's not that helpful for the way actual humans learn. Like we need time to think and wonder and ruminate and percolate and take that learning inward and downward and into ourselves. And I think of the work of people like Bernie Neville and other education lists. And they write about this kind of like sole perspective of education that the knowledge isn't just for an end goal of passing a test. The knowledge is becoming part of us and you can't separate the person from the knowledge they're learning. And that sort of over engineering efficient approach to education worries me. But I get why a lot of schools do it. So schools in England and I'm sure across the world, there are massive gaps in performance between different socioeconomic groups of pupils and huge pressure to try and close those gaps to give every young person the very best opportunity to go on and have a successful and happy life. So there's been a real push for this efficient catch up learning that closes that gap. And it's really well intended. But I think we're storing a problem to the future. So yeah, you might get these kids to get amazing examination results, whether it's at 16 or 18 for the use of AI. But so what? What have we created at the end of it? Young people that can only cope if learning is served up like perfectly to match their exact need at that exact moment. It makes me think a little bit about the difference between growing up watching analog TV and having like four channels and like dealing with it. And it's okay for like not to have your exact problem on the moment you want. And like watching my little girl now having like Netflix and Amazon and Disney. And like there's never a moment to possibly be bored. And there's never the idea of I'm just gonna watch things that happens to be on. And I think that's dangerous for it does to an individual's sense of being like, it's okay to be bored. It's okay things to be not perfectly efficient and served up to meet your exact need as a person. Like there's like a tolerance that we're losing by making everything so super personalized. And we're changing the expectation of what people think learning could and should be. We were talking before about just technology in general. Let's step back from AI. Let's talk about phones or social media or, you know there was even a funny image I saw with some Stanford professors way back in the day protesting scientific calculators, right? Because our technology there is coming because wrote memory of formulas was considered very important to the learning process. And so technology and education is not a new idea but these types of technology whether it's phones or social media or even now AI how are you seeing teachers and administrators especially deal with creating some frameworks or guardrails around what is allowed inside the school and also how to think about it outside the school. Cause once kids go home, their experiences are gonna be very different. And so you have this friction in the system. A parent may allow their kid to have a phone and social media and AI and some parents like myself will push back and say, hey you're not gonna have a phone until you're 16. You're definitely not touching social media until you're 16. And those rules don't seem to be, they seem to be at the parent level and then you go to the school district and those rules are all over the place. And so are you seeing some progress in defining ways to think about technology and education across a larger swath of schools? No, unfortunately not. However, one thing gives me maybe a glimmer of hope and that in the UK, and I'm sure that maybe this is the case in other countries we are very close to having every school being a smartphone free school. Like it was so close to it happening. Like it's not legislation, but it almost was. And the fact it almost was legislation is given every school, like feeling like the opportunity to say like, yeah, we're not doing this. And it's just helped them have that battle with any parents that thought maybe kids should have phones in school. So that gives me hope that we can do stuff on mass. We can make really sensible decisions as a sector, as a country that are in the best interest of our young people. But that's taken us 10 years to get to that point. Like 10 years, the information, okay, wasn't all there in 2015. It took us maybe a couple of years to realize the impact, but some deep critical thought back in 2015 could have mitigated some of the problems that our young people now face because of the way they interact with their phones. So maybe a bit of hope, like we can learn from the fact we can do stuff on mass, but I think for me, that the challenge is that we're never gonna get all parents on board. There'll always be a difference of opinion and we have to sort of respect that. I think what we can do is we can reach the young people. So we have to find time within our curricula to teach young people about what AI is, not how to use the AI. Like I don't think that's needed. Like everyone says they need to know how to use it for their future jobs. Like, well, let's just see, right? But what they do need to know about is what it is, how it's created, the ethical considerations around it, and think critically and deeply about it. So they're informed. Like, let's get them under the hood, get them that critical knowing eye, talk about the agendas that are at play here, why people are pushing it in certain directions and expose them to that. And I'm talking high school aged children, obviously not younger children probably. But the challenges for scores, like when? What are they gonna do that? The minute you wanna add in something to your curriculum, you're taking something else out and schools are already really struggling to get through the curriculum. And we've already had so many subjects up in arms saying, I don't have enough time for my subject anymore. And now we've got this huge burden in schools to teach young people this really important information and help them to think critically. And I'm just not sure where the time or space will come from. And also who's gonna teach them, right? So we need to teach the teachers first before they can teach the children or the staff. When's that gonna happen? How's that gonna happen? That is serious amounts of money that governments will need to invest in training teachers to do it. And I think this is the thing that annoys me the most about the whole past two years of AI. The only place you can get professional development about AI is, well, I say only, largely is from technological companies who fund it. So there's loads of free professional development out there. If you want it, surprise, surprise, who's running it? It's a technology company and they are not giving a balanced, nuanced, critical thinking approach. They're giving the, this is how you use it. They're giving the one side of ethics to make you feel like you've covered the ethics because there it is on the slide. And then they go into how to use it to make your life better. And it is really dangerous because there's nothing there as a counterbalance. And it's relying on individuals to go out, find the counter information, dig deeper and try and share it as much as they can on various platforms and networks that there are. You mentioned earlier, and I'm gonna get to actually meet you in person in the UK for a hackathon. What are ways companies and or even education can bring people along? You bring up a bunch of great points about the unintended consequences and really needing to understand and critically think about the technology, where it can be used, where it can be helpful, where it can be harmful, and to think about its implementation in a thoughtful way. But to ignore it seems not the right approach. And so what are ways that you're seeing school districts get their feet wet, where you're bringing teachers along and students along, and you're fostering those conversations before you get to the how, before you get to, oh, here's our new AI based, this product or that product. I think the hackathon approach has massive merit. And I'm trying to be conscious of my own and conscious bias here in terms of I ran many hackathons and sort of felt them to be very beneficial. But my senses and the feedback from doing them is people do find that approach to learning really useful when it's something that's new for all of us. So we're in this collective learning mode together. I think what I would do differently in yours, learn by iteration, is spend more time attending to there, making sure people understand how it is working and the ethical considerations before they get to experiment and play with it. But I think the experimental discovery with somebody who is pretty savvy alongside a sort of point out the potential side effect of doing too much of X or not doing enough of Y when you're using it is a helpful model. And I think it's scalable because it's a way of reaching people. Like you can, you do like a train the trainer approach. So rather than like one person having to go and like do this to all the schools, you get one person from each school to come to be trained by the trainer to go and run their own hackathons in their own schools. And you can cascade knowledge quite quickly in that sort of train the trainer method of doing it. As opposed to my concern is that there's quite a lot to learn to be able to do those sessions confidently and comfortably. And I'm just not sure schools have the resource to sort of give people that time to get that skilled up to do it is my big worry. But I think the other thing actually that could help here is that it's positioning, right? So like without doubt, I think Genitive AI is extremely useful in my role. Like I use it all the time and it's like a massive force multiplier. But I'm using it as like an executive assistant because I have really strong existing knowledge in my particular domain. So when I'm using it, I can check its output. And because I've got that really strong mental model, I can spot where it's wrong and I can edit and I can craft and I can ignore when I need to. I think if we were more upfront about saying that when we are talking about how to use it and not pretending it's something that a novice could use safely right now because it's not accurate enough, I think we could travel a lot further, a lot quicker in a much more safe way than trying to pretend it's equal use for everybody. And that's a challenge though, right? So how does school say, oh, you PIP and you James and you Joe and you Paul, you can all use chat ETP to help you because you're experienced experts. You can use it to help your lesson plans. But you Paula, Trudy, Susie, you can't because you're not good enough yet. You haven't proven you've got enough expertise to be able to check its output. Schools can't do that. So you either let it into your doors and accept teachers might be using it that haven't got the mental model to QA it properly or you ban it altogether, right? Because you can't have like a two tier system. And I don't know which one of those is the worst is the worst or the best option. How are teachers handling the flood of students using AI outside the classroom and then turning in homework? You know, a lot of learning is online. Take my girls, take your computer home and we try to supervise them the best we can. They don't have chat, GPT or any of the models on their school computers. But how are teachers navigating to understand, hey, has this student actually understood and learned versus copied and pasted something into a homework assignment? Yeah, like they can't. And that's one of the massive challenges. There was something to see the other day on social media about a particular homework platform that is just showed it has zero defense against AI chat pots just completing the work for children and is now a completely defunct website. And that's a surface that hundreds, if not thousands of schools are brought into using and invested in and trained staff and explained to parents and got kids on board and now they can't use it, right? So it's just gone. So that's like things like that is just, and the time and effort that means to those schools to put something else in its place. We basically have this whole challenge now as an education sector to think through how we are going to assess young people. People talk about the essay is now dead. Again, that breaks my heart to think that might be true for one moment because for me, the essay is this beautiful form where we have to wrestle hard with our thoughts and we have that desirable difficulty and the grapple and the craft and the refine and the reflect. And it's not just about the end product you produce for an essay, it's a process you go through. It deepens our learning. It makes us more cognitively able and fulfills us and evolves us as humans. Now that's potentially dead because of genitive AI because we can't check whether people have cheated or not. So we train teachers. One of the main parts of my role is we work with novice teachers. Part of their process of becoming a teacher is they have to write a couple of academic essays that sort of master's level. Like we're in like a virtual crisis. Like what do we do? Like the checkers can't detect the AI reliably enough. So we turn them off because we just can't rely on them. So people could just be full on cheating and we just have to rely on the fact they're not because it's not professional. If they did that and we talked to them about the ethics of please don't cheat and it's not good for your cognition and it's not good for the integrity of the program. But literally we have no way of really checking anymore. Like we had a really great assessment method, the essay and it's just gone. And I know hundreds of universities are grappling with this and maybe it's an opportunity for something great to come out like maybe there's new, innovative ways that we can assess people's learning. But at the moment it's a real challenge. I always subscribe to writing is thinking. Yeah. I was on a round when I wrote my first book and going through that process to your point, it was hard. It was extremely hard to sit down and put your thoughts into a narrative structure and write prose and really think about what you were trying to say, grappling with your brain into the written word and trying to communicate to your point about advertising. Like trying to persuade someone that you have an idea that's interesting enough that'll help them think. And I walked away from that experience, one proud of the fact that I went through it, proud that I struggled, which I think was the point of the exercise. I mean, yes, I got my thoughts out into the world and yes, the book has helped some people. And now I'm starting to work on my second book. And I'm at this crossroads, just what you were saying is, hey, here's this technology that I know can be helpful, but I don't want it to be efficient because if I make it efficient, it makes it less impactful, probably to the reader, but it makes it less impactful to me. I do this podcast, I do my content because I'm curious. I want to learn and I'm willing to invest my time in talking to people like you and writing books and articles and stuff because it helps my thinking and it helps me try to navigate what the future of my career might be or how to raise my kids. And the push for efficiency, as if it is the end goal, I think is the thing that I struggle the most with. Absolutely, I really agree, but we have this commodified education system which is all about exam results to get on to the next step, to the next step, to the next step. And this idea of loving, learning for its own sake and what it does for us as humans has been somewhat squashed in the current educational systems in the world. So it was almost like we had the perfect conditions for AI to sort of just plop in nicely to this pretty poor system we had because it's just a pretty much commodified system anyway. So this is like the rallying call. This is maybe the chance to sort of say, look, this is gonna push us to an efficiency extreme that none of us are comfortable with now. This is the chance to rethink what we want our young people to experience in school. And we go back to the origins of the word school and I can't remember the exact etymology, but the idea of it links us to think about space and free space and the importance of that in learning. Like you can't rush some learning, it's organic and it isn't in a lovely straight, ever-increasing, neat line. It's up and down and it's messy and that's great. That's a great part of being human, messy learning and that sense of achievement when you finally crack something. An example that comes to mind for teaching is that lots of teachers are using AI to create lesson plans and I totally get it. Like it's quick and efficient and teachers are overworked and you're knackered and it's 10 o'clock at night and why wouldn't you just quickly knock up a lesson plan? But the danger in doing that is when you come to teach that lesson, if you haven't been like a true architect in that sequence of learning, you're never as close to it as you would have been if you'd architectured it yourself and therefore I think the quality with which you teach is then lessened. So the challenge is we need to find ways to make teachers less tired and have more time, but it's not through making them think less by giving them AI as a sort of a supplant. It's about trying to reduce the number of hours they have to teach by being more creative in timetabling and budgets and finance. I don't want them to stop thinking as hard. Like that's what makes them great teachers. And now you also have this real concern that like what do we do for those people that are new to the profession and don't know how to plan a lesson? It's not about the end product as the process they go through, right? So if they turn straight to the AI to help them create that lesson plan, they are never gonna build up the mental model to be able to quality assure the output. They are never gonna learn from the cells how to do that. So they will become passive receivers of the AI produced content for the rest of their teaching lives. And this will sound dramatic, but there will come a point when all the teachers that learnt to lesson plan without using AI are dead and there's only the ones that are now sort of like probably in their 60s that have been brought up using AI to plan lessons. And then we are totally in that passive state of we're just relying on the tech to give us our intelligence. We've handed it over to them completely. You said something that was interesting around the process of learning being messy, ups and downs and non-linear and non-efficient. That's life, life itself. One of the things that I experienced and I'm sure you did to me, others, is that school is about what you learn about algebra and history and science. But it's also the life lessons you learn through the struggles of failing a test and studying for a test and thinking through things and not understanding something. I was actually doing homework with my daughter this morning and her struggle of learning how to do this math problem was her learning how to learn. It had absolutely nothing to do with the math problem. The pride that she felt at the end of this homework assignment we were working on was that she learned how to learn something. And it was messy and she was upset and she was crying and she didn't feel good at the beginning of it because it's really hard, but we stuck with it. And when you take out to your point, when it all becomes passive and it's all efficient and there's no more struggle and the answers are just there for you, I wonder if it impacts the students as they get into the real world because the real world doesn't mimic a passive AI technological system. There's other humans at least today that you have to interact with. There's other systems and organizations and all sorts of complexity, not to mention the human brain and your continued fight to keep your thoughts together and wake up every day. I wonder if losing that is one of the major downsides. I think it is and there is decades and decades of research on the link between metacognitive ability, so the ability to think about the way you think or learn about how you learn and your outcomes as a learner. And we know that you can teach people to be more metacognitively able than they perhaps naturally are. It's a teachable skill. Like, oh, we know this and there's decades of evidence around this. And what's really interesting is some of the research I've read recently around students using, this was using chatGTP. It was a study that was done only in 2024, say last year, by two researchers called Hueng and Chui. And they had looked at 766 different research papers that all evaluated the impact of using chatGTP to support learning. And this was, admittedly, this is university-age students because that's where most of the data exists at the moment. And what they found was, surprise, surprise, chatGTP definitely helped to improve their short-term understanding of the topics they were studying. And it particularly helped those learners that were identified as struggling learners. However, the studies also showed, and this is 766 different studies, that it limited the development of those higher-order thinking skills compared with the more traditional learning methods. And it posed some significant risks in terms of overreliance. And those higher-order thinking skills that link to the ability to be metacognitively able, we know of it all to survive as humans, to be healthy, happy humans that can navigate a very complicated, messy world. And if we're denying them our young people of the opportunity for the sake of quick, efficient learning and immediate good grade outcomes, we're actually probably doing them the biggest disservice we possibly can. What would you say to someone who's listening that has children? A lot of us are going through with phones right now, or social media, and I mentioned earlier about the anxious generation and how that book kind of frames it and takes you on a journey, helps explains it and then provides some rules. But for parents out there where they feel, hey, I wanna give my child the best opportunity to succeed in school, right? I want them to get good grades in middle school. I want them to get good grades in high school. I want them to go to a good college because the data still proves that if you are a college graduate and you're educated, your financial outcomes, health outcomes, and a lot of things are just net better. As you, as a parent, work through this, what is advice that you would give someone today, given that we're in this messy space where the technology is not mature, where the tech optimist will hype cycle it and tell you that it's perfectly fine without having guardrails and all of the things that we just discussed? I think that the simplest piece of advice I would say to people is to take the pressure of yourselves. Like, we can pump the brakes on this a bit. Like, there isn't this burning urgency that the hype is trying to make us all feel. Like, we have got time. We don't need to make rash decisions about this. We can wait for some stronger evidence to emerge about some of the consequences of this. And that fear that your child might be left behind or fall behind, I would say, to have confidence that that won't happen because we've had a really good way of teaching young people for a long, long time. And it largely works. Not for everybody. I know that. I know there are some young people that the traditional learning methods don't work for. We need to talk about that on a whole separate podcast properly. But I think that idea of just slowing down and giving time for just to settle and seeing how things develop is certainly what I use with me to guide my decisions. I also think I accept that there could be a place for very safe, protected use of particular AI applications that have been developed in very tight tandem with schools that attend very closely to their safeguard rails and to what we know about cognitive science. I've yet to see them be developed and maybe I just haven't looked hard enough. But I am confident that there's some good person out there doing that kind of work. So there could be some opportunity. And I also, by the way, totally get that. I remember teaching in schools, it's Thursday or Friday, it's period five. You're knackered as a teacher, you're on your knees emotionally and physically, the kids have had enough. The quality of your teaching for that last hour of the day might be shocking. Like you can't teach five outstanding lessons every single day. We're just human, it doesn't happen. So I can see how a school, a head teacher might turn around and say, you know what, for that last hour of the day when the teacher's knackered and the kids are a bit knackered, I'm gonna give them a device and they're gonna have an AI tutor, take them through a series of consolidation questions of stuff they've already learned just to reinforce their learning. And it gives my teacher a break and the kids really enjoy it and it gives us some quite good data as well. Like I can see an argument for that. And it's appealing and it's strong. My only worry is, do we really believe we can keep it just to that one hour? Or does it begin the creep? Like we start to normalize it. Well, if you can use it for one hour, why not use it for two hours a day? Or one at three hours a day because, hey, I'm saving a fortune how much I'm having to pay my teachers and the kids are all really quiet and well behaved because they're just locked into their little device and their headphones. And I'm just worried that we couldn't stop the creep because the agenda and the tech companies will be big to push this for greater and greater use. I think that's the motivation basically. So that's what worries me is that, yes, there's massive potential but can we put in enough guardrails or rainstakes? I'm not sure. So as a parent for the time being, I would say, I would just take a step back and just let's see what some evidence tells us in a few years time. The one thing you said that I work very hard with my kids is to ensure that they're bored. And I know that sounds crazy and controversial. I agree. But like the- I totally agree. The pressure of Netflix and Disney and the iPads and all of the technologies which are literally designed to engage you. The algorithms are a lot more sophisticated than our ability to turn them off in many cases. But to go outside, to turn it off and to spend some time being bored. Yeah, I couldn't agree more. I think I say it to my daughter says to me all the time, I'm bored, mummy. And I always say, nice to meet you bored. And she hates it. It's like a stupid joke, but it's really, really important that you be bored. Like it's really like a PR campaign to change the connotations of the word bored from being negative to being something we should celebrate. I would love chance to be bored. It doesn't happen in our lives as adults now really, especially trying to grapple with this AI stuff which definitely wasn't in anyone, in my world's job description two years ago, but somehow we're trying to make space and time for it. So I think being bored is a great thing. And I also think it ties into this idea that it probably sounds a bit cruel this phrase, but I think a healthy level of disinterest in your own children is to be encouraged, right? And that's a bit of an extreme way of putting it. But this like over-helicopter-ing parenting style where we need to know every single thing. We've got apps that tell us exactly what they did at their school and how they got on that club. And like, I'm not sure it's healthy for young people to be so central in their adult, in their carers lives. I think it's good. Like they need to feel loved and cared for. And I'm not suggesting that isn't the case for a minute, but I think there is something about, we don't need to attend to their needs every single moment of the day and make sure they're super stimulated. And we know every single thing they've done. We need to give them time to be by themselves in a safe environment, of course. And that inwards learning process that goes on, that slow cogitation of ideas and space to think like, we know that's good for young people. And we seem to be robbing in with those opportunities because there's a need to keep them stimulated and busy and using every opportunity to excel and grow. And it's like, it's race, a race for what? Like to get into the best university and get the best job, but be a really unhappy human at the end of it because you're really anxious or really stressed or have no resilience. Like, I don't know what this end game is we're trying to chase down. Pip, I couldn't agree with you more as someone who is in the throes of trying to raise two children and also trying to get my own relationship with AI into a place where I understand it, given the hype cycle and especially my history and working in big tech. I'm gonna put a lot of the research that you have provided in the show notes. And I encourage people to check that out. As we navigate the complex intersection of technology and education, Pip reminds us the most important question isn't how we use AI in school, but what kind of humans we're trying to nurture. Pip, thanks a ton for being on the podcast and taking time. I know it's a little late there in the UK. And thank you everyone for taking the time to download this episode and listen, don't forget to subscribe. And until next time, question the hype, trust your expertise and most of all, stay curious.