#332: Most People Are Using AI Wrong, Are you?

Hello and welcome to the EdTech podcast and this week we're asking how well do you know how to use AI? For many of us it's become a faster search engine, ask a question, receive an answer and move on. But today's guest argues that this is not only the weakest way to use generative AI, but it also may prevent us from developing the skills that we need for the future. Joshua is the co-founder and CEO of Mindstone, working with organizations around the world to help non-technical people transform how they work with AI.

Before that, he co-founded Super Awesome, developing technology designed to keep children safe online before the company was acquired by the well-known Epic Games. In this episode, we explore why AI transformation is fundamentally a people challenge, not a technology project, and why asking AI to question you may be more valuable than asking it for answers? And why schools that simply ban these tools may be encouraging the very bad behaviours that they're actually trying to prevent?

We also challenge some widely shared assumptions, including the claim that every AI query consumes a high quantity of water. My own misconception that it's one bottle per search? And are the environmental statistics that we repeat giving us the full picture? Or has the conversation become more complicated than the headlines suggest.

From safeguarding and skills, inequality to agentic AI, job description and the changing purpose of education, this is an honest conversation about what happens when the way we learn and work changes faster than the systems around us. Welcome to the EdTech Podcast. Hello and welcome Joshua to the EdTech podcast. It is a delight to have you here.

We're going to be talking today a lot about AI, but please do tell our international audience and listeners a little bit about who you are, please. Yes, so I am Josh. I'm co-founder and CEO at Mindstom, and we do AI transformation for mostly non-technical people. Some of the biggest companies in the world, Home Depot, Hyatt, Delta Airlines, Pearson, Fortnum Mason, across the board.

And we help people understand how they can use generative AI, so Chat2BT, Claude, Copilot, in their job on a daily basis so that they can do either more or better work in less time. Incredible, which everybody must love to hear, right? Right now, it does seem like it, yes. Amazing.

And so just to backtrack a little bit, tell us a bit about some of the things that you've done before this because I know you've got an absolute wealth of experience especially with startups and companies and innovation generally how did you get to to this point well so I'm originally Dutch did about 13 years there then moved to Switzerland and I've been in London for about 16 years now started my first company when I was 16 basically building well at the time websites not really apps for friends of my parents and then friends of their friends and that grew into 30 high school students building apps for private banks across Geneva, which was an interesting time of my life.

I was the very awkward high school student turning up in a suit at high school because I had a meeting in a bank right after, which didn't make me the most popular place at a high school for sure. Wow. I mean, I bet looking back there, people are probably a bit more like, I really wish I'd been more like Joshua. It was fun.

It was definitely character building. And so then when I moved to London, I started another company, a company called Targets, which actually went nowhere, but it was every buzzword in a single sentence. So we were doing location-based, real-time community marketplace. Wow.

I know, every buzzword you could think of in a single sentence. So the idea was that two different people in the same place at the same time that are looking for complementary things, could you match them together? So two people at the same Starbucks, like one looking for a laptop, another selling a secondhand laptop. Can you match them?

Yeah. We learned the hard lesson. We had some investment, we got an investment offer, but they wanted a lot of the company. And I thought we could get better.

I declined it. And then the company just went under entirely because we were at the end of university and people started getting job offers. And so we weren't fast enough. So that kind of never really took off.

But that was then also the start of Super Awesome, which was the last company I built. I was co-founder, chief product and technology officer. And that became one of the biggest kids technology companies in the world. So we built a whole bunch of tech to keep kids safe online.

So there's two kind of big elements to it. One was, if there are ads in games, how do you make sure that the data of kids is not being collected? And how do you make sure that there are ads? These are ads that are appropriate to the kids that are actually in the game.

So we were working with Disney, Lego, Hasbro, Mattel, every kid's brand you can think of. And on the other side, we also built all of the kind of verified parental consent stack. So I think basically any app today that you would use that is for under 13s that would ask for parental permission, that will go through our tech in one way or another. It was ultimately acquired by Epic Games, the creators of Fortnite.

Amazing. I mean, already I'm absolutely in awe of the things that you've been able to do, things that have really transformed systems, education, gaming, safety of our children, which is phenomenal. You must be very proud of the things that you've been able to develop. It's, yeah, I've had really good times and I've had a lot of privilege to be able to see behind the scenes of how messy everything actually is.

It looks very nice from the outside where you talk about it backwards. It feels extremely messy on the inside when you actually do it. Yeah, I can absolutely imagine. But some of those things, I mean, especially when we think about the world that we're in now, where, you know, government entities all over the world are finally putting in restrictions and safeguards and really considering how they're responsible for the parental control, but also the children's use of technology and how we're doing that.

And actually, that sounds like it's something that was right at the precipice of that, the real forefront of safeguarding. Yeah, absolutely. For what it's worth, it's an interesting one, which is as much as I spent 10 years of my life building that, I wish we had done more of that then because I'm not the biggest fan of now the government coming in and basically removing parental responsibility whilst doing it. I would have been a much bigger fan of pushing what we had already built and just making sure the parents have the tools and the knowledge to be able to make the right cause in the right time rather than just blanket banning everything and making it laws.

Absolutely. And I totally agree. I'm someone who's been trying to support parents in understanding how simple it is to use some of these tools because it feels so far removed, doesn't it, from that ability to be able to allow your child to use something but also keep them safe. and there is such a disconnect within that information and that space.

I think even with the laws, there's going to be so much work to be done. There was an interesting interview with a parent from Australia when the Australian ban came out who basically didn't really mind if their child kind of superseded all of the rules. And unfortunately, we are also having to deal with that side of things as well, aren't we? So amazing.

I feel like I really want to ask you, so imagine at school, what were your teachers saying about you? You're this entrepreneur, You're a kind of a unicorn, if you like, of someone who can really see innovation happening and very driven at a very young age. Yeah. So I guess it depends on when they would have asked because this, like the entrepreneurial bit became more apparent.

I guess it was fairly early. I actually forgot I had a paper route when I was 12 in the Netherlands. So that was, I guess, a starting point. I always had a lot of energy, so a lot of sports, and I was definitely not the easy child in the classroom.

So I was both ends of the spectrum, I think, which is when I had always had the questions, but I would also not be satisfied if the answer didn't make sense. Which is fair enough. But that did put me into a few interesting situations because sometimes you have the really good teachers incentivize that. really want to go deep.

And then from time to time, you'll have teachers who kind of, they give you an answer. And if you don't like the answer, then they're just like, well, that is the answer. And if you don't, if you don't agree with that, then that gets you into trouble sometimes. That whole, because I said so.

Yes. Amazing. And I guess where you grew up, did you, in terms of education, do you feel like there was, it was very flexible for you to be able to kind of curate your own skills? Yeah, so my dad is a pastor and my mom is a philosopher and they could never agree on the type of schools that I should be in.

So by the age of 13, I had done six different schools and six different school types. Wow. But I was constantly moving schools. And I guess that did give me a special perspective on different ways of approaching.

I jumped to grade ones and then I got held back grade ones. And it was just, I went into like a free school and then it was more like a Montessori and then it was else. So it was interesting to see the different approaches to it. But I definitely felt like I had ownership to a degree of being able to shape.

And I was able to follow. So I didn't have trouble, intellectual trouble kind of in school. And at home, I was very incentivized to go and do my own thing and pursue what I wanted to pursue, for sure. Incredible.

And I just feel like there's definite kind of elements to the way that our journeys create and move that make us who we are. And it feels very much that at such an early age that you were able to have that autonomy over. And you said, you know, you're kind of an agency over that. And that's amazing.

And to now have Mindstone. So tell us a little bit about that. So I came across an article linking to the Tony Blair Foundation. But tell us a little bit more, a bit more in depth about what that does.

Yes. So when we're in the process of selling Super Awesome to Epic Games, Great name, by the way. Well, when we got acquired, we had a funny conversation. One of the first things was with Tim Sweeney and us.

And we were kind of going back and forth. Well, who has the best name, Epic or Super Awesome? Which was, he obviously thought it was Epic. But yeah, it was ultimately them who acquired us.

So I guess he wins. When that happened, I really went through an interesting period of trying to figure out where did I want to spend the rest of my life in terms of problem solving space. I've always been a big believer of kind of compounding your knowledge over time. And so I do think there is some merit in going deep and surrounding yourself with people that are in a similar space for quite a while.

And so I explored, I wanted to really have an impact. I explored both healthcare and education. And healthcare, about one week into exploring, I realized that it was going to take me at least five years before I could be remotely helpful because I just didn't understand enough of the space to start to think where could I be useful. With education, I felt because of my kind of weird past around it, I mean, I did my MBA remotely as well.

So I had remote learning. I had kind of different schooling and the whole thing. I felt that I had stronger opinions earlier on that had merit. And so I decided that education in general was probably the space I wanted to spend more time in And then yeah started Mindstone on initially kind of three big pillars One was I read a book called How We Learn by Benedict Cary, which was really about what do we know about how the brain works and how that relates to what we retain, how we retain, how we learn, and how little people actually are aware of it.

And so that was the actual starting point. Like, wait, we know all these things about how we learn. Nobody else seems to, or very few other people seem to actually learn it or apply it. Well, there is a business to be created here.

I was talking about it as information arbitrage at the time. It's like, wait, if you know something and nobody else does, then you can bridge the gap. Yeah, amazing. So that was the starting point.

And then I also had this thing where I always loved learning, but I hated education. I really didn't like the schooling system. And I thought, okay, well, there has to be a better way to go and do that. And the average time, already at the time, so this is now six and a half years ago, the average half-life of a skill was just going down and down and down and was getting close to about four and a half years at the time, I think it was, which meant it was about to go under the time it would take to get a degree at university.

And it was like, okay, well, if the half-life of a skill is under the time it takes to get a degree, then why are you getting a degree? Getting a degree. And so that was the other thing. This system is about to collapse.

Can you build something in parallel to try and be more helpful and be more future-proof? And we started building that. And then three years in, Chachupiti happened. So everything in relation to learning suddenly changed, both in terms of what you could build, but even more importantly, in relation to the skills that would even be relevant in the future.

And I think we are going through, from that perspective, probably the most fundamental shift that we have gone through in hundreds of years in relation to skills that will be valued in the future. And so we just decided to focus the entire company on helping people understand how to use this technology, understand how to relate to this technology in their lives, in their work. And that's where we have focused all of our time since. Amazing.

And I love that, the idea of being able to relate it, because I think that's one of the big struggles for people with regardless of AI, but generally digital transformation and digitalization of anything is where do I sit within this? There's that big fear factor that some people love to kind of push, which is it's going to take over and your job will go and everything will change and you won't know how to do it unless you know that kind of really pushing people into the depths of worry.

But actually, if we can relate that to something that can help people to feel secure and belong and understand, then that's a real change mechanism. How do you do that? So it's incredibly difficult because, and there's a few elements involved. The technology is so, appears to be so easy to use, that it's extremely easy to think you know how to use it.

And it's one of the biggest problems because the amount of people that go around, say, really know how to use the technology because they use it multiple times a day is really staggering. The number of people that actually know how to use it is literally less than 1% at this point, I would say. And I'm probably still being generous there. Now, most of this is due to the fact that the technology itself is not actually the thing that helps you get to a behavior change.

You have to change the way that you approach a problem or the way that you approach a particular task. And if all you're doing is executing the task in the same way you always have and trying to shoehorn AI in there one way or another, you just never get to the actual return or the value that AI can offer in the first place. It's kind of like, imagine you are two people working on a problem, but you don't change the way that you approach the task as if you are the only person working on it.

You're not going to get any value out of the collaboration. You're approaching the task as if you're the only one doing it. And that is something that is really difficult for people. I mean, in many cases, you're trying to undo 20, 30, 40 years of habits.

And so there's a few things that we start with, but a few techniques or a few small things, once you know about them, really change the way that you think about the technology. And so it's kind of the starting point. One of the biggest ones is everyone right now, or most people right now, would use generative AI almost as a search engine replacement. A lot of people go around, yeah, I know how to use generative AI.

It's basically replaced my Google usage. Well, that to us then means, okay, that means you don't know how to use AI. Because if you're using it as a search replacement, it's actually the worst use. Yeah, the worst.

and also the most unsustainable my fear for every time someone said that to me I'm like do you know how much water it uses for you to ask that question and you could have just used Google and it would have still used AI in a very clever way but it would have been less energy yeah so I would like to address the energy and water usage question for a second because that is right now quite a bit of misinformation actually is it? yeah so a single almond the production of a single element would get you about 30,000 queries on chat GPT.

So if we're talking about actual water usage, let's make sure every time someone eats an almond somewhere, we realize that is 30,000 queries that you can go and do on chat GPT. So we have to be fairly clear about what it means. And then the same time at an energy level, The energy usage is definitely increasing, but the entire world's data centers combined are less than 2% of the world's energy. And that includes AI, which is less than 10% of that.

Let's also be clear that we're talking about energy increase that's exploding, but it's less than a percent. It's less than half a percent of the world's energy. So when we talk about the energy in that way, it is absolutely increasing. increasing.

And so if we talk about, hey, what is the trend and where do we go in 10 years from now? We really have to be, we have to watch out and make sure that we plan for that. But to talk about how right now it's consuming that much energy, that is not quite true either. I see.

So that was interesting because I was told that by somebody who lives very close to one of the Google data centers who has power cuts when they do their updates because of the amount of energy it draws from the grid. I'm also really intrigued now how much water it takes to grow an almond. I know, I know. It's one of those things, yeah.

But that's amazing. And I think this is, you know, these are the things that we get told and we talk about and is out there. And when we're talking about sustainability curriculums and we're looking at those different elements is vital, but there's so much information around. And I guess this is one of the problems with AI, isn't it?

Is that what is real and what is not? And unless you're at the data source with the research, that's really hard to understand and know. Amazing. So, okay, so let's come back from understanding a little bit more about the energy use, thank you, the skills and requirements of people and human users.

Yes. So, yeah, I was talking about the kind of the main or the starting point. One of the biggest ways to remove many of the risks of AI and change the way that you use it is actually asking the AI to ask you questions rather than getting it to answer questions. So, So you have a problem, you have a task or a thing that you need to start with.

Explain the problem to JantyBT or Copilot, whatever you're using, and then ask it. Okay, now ask me one question at a time, waiting for my answer in between, to help me think through this problem, to help me discover angles I haven't thought of, to help me think of the blind spots I'm not seeing. And suddenly you're using the AI to enhance your thinking rather than to outsource it. And it's so interesting how few people think about it.

It's an extremely simple thing. Once they know it and they start using it, it's like, okay, wow, this changes entirely the way that I use AI. But very few people would naturally get to that point. It's like we don't ask the machines to ask us questions.

They have always answered the questions that we ask. So that is kind of one of the big things. The other is this idea that you approach the conversation with AI as a conversation. One of the controversial things we say is that it's the more the more helpful framework for you to understand how to get the most of AI is think about it as a conversation with another person.

doesn't mean you kind of sidestep the whole personification thing, like everything else. But it means that everything that's true when you're having a conversation with someone else is likely true to also be the case when you're having a conversation with AI. So one of the things, if you ask a question or the first answer that comes back isn't great, that doesn't mean that AI can't help you. It's like when you're having a conversation with another person and the first answer that comes back isn't great, you don't stop the conversation.

Sorry, you're no longer worth having a conversation. I'm just walking away. I'm sure there are people out there that do, but I hope I totally understand where you're going from. The number of people that do that with AI, they get the first thing back, oh, that was useless, and then they move on.

Yeah, it can't help me. It can help other people, it can't help me. Exactly. And instead, if you'd given it two or three tries, you'd just, hey, that was not what I was looking for.

Like, this is not, I'm actually more interested in X, Y, Z, or whatever. Like, the second or the third answer That might have been the actual one where you get the real value. So those two bits already change fundamentally. Once people get those two, you're on a path now to start to explore much more.

And then you can start to layer in everything else. All the way to when you start to think about kind of agentic AI. And the AI can start to action things for you. It is a journey because it's hard, again, to change your habits.

and to understand just how far you can push the AI and how you do it. There is a bit of building up your system over time. Similar to, and very few people can relate to it, but there is an existing paradigm when you have like a really great chief of staff. Some people have chiefs of staff that they work with for years and years and years.

Once a great chief of staff has all of your context, they know your personal life, your professional life. They know how you make decisions, what you like, what you don't like. The more they know you, the more they're able to start taking work off your plate and actually do that in a good way. The same is true for AI.

And so the more context you can give it and the more context it builds up over time, the more useful it can become. And this is one of those things I think that is important to understand. And it's a bit of a question mark on how the kind of inequality of the usage of AI is growing as well, which is that the people that have been building that context are able to get dramatically better results from AI today than the people that haven't. And I don't think it is that easy to instantly build that context, which means that it is something that compounds.

So I've now had three years of building different sets of context for AI, which means that even if someone had the same level of skill as I would have to use the AI, because they don't have the system built up that lets AI help understand what I'm trying to do, they wouldn't be able to get to the same result. And so that's something that I think is special about this particular period we're going through, because it does make me question, when do you erase that inequality? Like it's building up at the moment, but it's like if somebody only starts using AI in a year from now, are they permanently going to be behind because they're just catching up trying to build their system?

Or is there going to be a good way to kind of bridge that gap faster? I hope so, but right now we're seeing the opposite happening. I was going to say because at the moment we are seeing just like many things in education across the world is actually that disparity and that inequity and access that we see and I wonder when you sort of kind of showing that slope is is my also thinking is those people who are are young learners who yes we we've got lots of students and children and people globally who are on it who are most definitely under 13 and and using AI apps that I'm sure but actually when they start to use it in certain ways, it's actually that building up as well.

So they're going to be starting from a certain point, perhaps having more understanding of what AI is because they've grown up around the idea of it and understanding that. But they will also be starting to some extent from ground zero. Yes, absolutely. Now, I'd say the problem for younger people is a slightly different one as well, because they're struggling with many, if not most schools at this point, just have a terrible AI policy, if they have one at all.

Most of the time, it's just plain ban, which doesn't mean that kids won't use it. It just means they use it in the worst possible way, which is basically... In secret. Yeah.

In secret. And they do exactly the thing you shouldn't be doing, which is they outsource their thinking and they're just going to submit different pieces of work that they haven't actually thought through themselves. And that is the worst. I'm a big fan of what Ethan Mollick has done.

I know it's not in kind of early education, but when ChatTBT came out, the first thing he did is he mandated the use of ChatTBT in his classroom. He literally did the opposite. And I think it is a dramatically better starting point, which is this technology exists. Just like you're going to be using a pen, and I'm going to assume you're using a pen to try and do your work, I'm going to assume that you're going to use the best tools available, including ChattoBT, which means you're going to have to show me how you have improved on that work, how you got to the result.

And if you do worse than what ChattoBT does out of the box, you have failed by default. The whole question becomes, can you show me how you have taken this piece of work that is now the baseline, which everyone can get within a second, and you have iterated on it in a way that is constructive. And it's just such a, I wish our school system would just take that as a baseline. I'm being directive of like, how do you then figure out the issue?

But at least we should all agree that the baseline is just that we have access to this. Yeah. And I think that idea of removing the negative connotations and this idea that it is wrong and it's bad and it's cheating is one of the most important parts of that. Because as you say, they're using it anyway, they're going to be then using it under cloak and dagger.

And so actually teaching them that we're aware that we know how to use it, that we can use it as a trusted tool, that we can be clever with it, we can use it as a coach. And, you know, even earlier when you were talking about the questions, I think one of the things that I often think, well, we see it, we go from primary school where we ask hundreds of questions a day, we move into secondary school where we reduce that to 20% of that, if that at all, as a learner in the classroom to listen to what I'm telling you, don't ask me too many questions because, you know, I'm going to be the one that's right and you're not.

And you are needing to absorb knowledge. Well, we might as well put them in front of an AI bot and say, here's the information, read it. In some respects, that coach, that ability, and we are seeing some really innovative schools saying, use it as a coach. It's going to ask you questions.

It's going to draw things out from you. It's going to try and challenge you. is a much better way of developing and thinking and helping our young learners become critical and become human-centered users of technology. And that's really, I couldn't agree more, is where we need to go with it.

And actually that allowing, and many schools, as you say, I think blanket ban of anything is never a good idea because it doesn't instill the right messages, but we need to upskill everybody. And you mentioned at the beginning upskilling and kind of integrating. How does this help you work better, smarter, faster? So how do you do that at Mindstone?

What's the process of helping companies, helping education institutes do that? So we think about transformation as including many different aspects. One of the things I've had to change my mind on recently, so when we started Mindstone, I really thought, because as I said, I always loved learning myself, I thought our goal or our task was to like light a fire, get everyone to really love learning. Because if everyone loved learning, then that would be easy, right?

It was a very idealistic way of looking at things. As I saw everything unfold currently, which is happening with AI, I did end up changing my mind as we built MindStone, just because not everybody loves learning. Doesn't mean they shouldn't do it. Everyone has to do it in one way or another.

But it's totally fine for some people to not enjoy it, not enjoy the uncomfortable process of having to modify how your brain works, how you act in different ways. And so when we do this, we really look at the full transformation picture. We often start with the executive team first, actually almost always, because right now the people that are in charge are just as clueless about this as the people that aren't. And the problem is they think for most changes in technology today, actually almost everything up until today, with the exception maybe of the introduction of the computer, the executives would delegate how this new technology would be implemented throughout the organization.

So it often at the moment sits on the desk of the chief technology officer because it is tech. The first thing we do is we do a session with executives where we help them understand how they should use the technology themselves. Because it is a knowledge enhancer or knowledge work enhancer to a degree. If there's anyone in the company that has a huge lever on every decision they make, it's them.

And so they should be the first users of it. And if they don't, it sets a message to everybody else. And it also means that they're approaching it in the wrong way. So the first task that we kind of set ourselves is that the responsibility of rolling out AI needs to move from the chief technology officer to the chief people officer.

Right. Because it is a people problem. It is not a technology problem. And why is it so easy for people to think they know how to use the technology?

Because the technology use, rolling it out and getting people to write a message in a text box is not hard. The hard bit is the change management, is the people changing the way that they work, the way that they go about particular tasks. So we start with the executives. Once they understand how this technology and why this technology is so different, then we start to work with the rest of the organization.

And in our case, we kind of built a hybrid approach. We do live cohorts. And this is mostly because, I mean, we try to do entirely asynchronous, kind of just everyone self learns. The software engineer in me and the company builder in me was like, well, that will scale infinitely.

So why don't we just do that? Didn't get any impact. If it's just people looking at a screen, people just don't get it. Like that's why when people go through a Coursera or a LinkedIn learning course, like it's great, maybe you get the completion rate and that's even one thing.

But nobody actually absorbs the skills. They don't actually change the way that they work. So we have one element, which is we do these live demos combined with a learning platform that is entirely personalized to each individual. So the idea is the live demos are trying to convince people that it's real.

We're going to just show you how you can use this technology. Sometimes a live demo fails, but that's actually the interesting bit. When we fail with the live demos, it tends to have a bigger impact, a positive impact. Rather than when the live demo is all slick and everything goes right.

Because people are like, I'm being tricked here. It's not this easy. The live demo is so important because there's a whole narrative at the moment about is AI hype or not? Can it do the thing?

You've got a whole people thinking there's snake oil salesmen going around and saying it can do everything, which to a degree is true. There's lots of people saying everything. But the live demo is kind of one piece. And then you have, if you're convinced, okay, now maybe there's a bit more to this than I thought.

And then you get an actual practical application that you can directly apply in your job today. and it's relevant to you, it's based on your context, your role, your work, that is when the penny drops. And then once they've got that first step, then it becomes about, can we sustain momentum? So can there be a heartbeat throughout the organization where as you do work, you have enough signals on a daily basis that just reminds you of, wait, wait, maybe I should think about how AI could help me with this thing.

And then once you've got the ball rolling, then the rest will come. so long as you take the heartbeat of life. Amazing. And do you find that in organizations it's that very classic kind of, you know, the first kind of cohorts of people are very enthusiastic and they're going to embed and they're really spearheading it.

And then you do have your sort of laggards who are trailing behind and still very cautious at the end of all those sessions. Yes, it does exist, but you can temper that. So we do a lot of work. And part of the beauty of kind of life cohorts is we will select a few champions in the cohort whose responsibility it is to shout out the big wins and help the laggards not become too laggard.

And so every week they will get a report on how things are going and where they should put their attention. We definitely do see it in big corporate where people are just, when the organization allows for a space to actually properly lean back and just not engage, there are people that choose that option. We do also see the complete opposite. We've seen companies, hundreds of people at a time now, who just jump in with both feet and the entire organization starts using it.

When you see enough examples around you of people who are getting the real wins, who are actually getting many hours a week back, who finally are getting to work on the things they always wanted to dedicate time to, but the admin work came in between that. It is when used in the right way, it truly can help you do much better work and work that you might enjoy doing more than before. Yeah. And it leaves that opportunity for the things that you really enjoy doing more.

I think that's for me is the bit that I really love, especially as a kind of independent business person. And I like the fact that I can offset the things that I don't enjoy as much to be able to be quicker with those so that I can spend more time on those really those bits of my job that I absolutely love. And I guess that is a real win for people in terms of change is if they see that that's a real positive, isn't it? And so if we come back to some of the work that you did previously with Super Awesome, the idea of the safeguarding, the child safety and how are we seeing that?

You know, we've spoken a little bit about people thinking they know how to use AI, but actually a lot of that isn't necessarily true. When we're looking at our younger people and we are putting a lot of information again, you've spoken quite a lot about the idea of we have to tell it and feed it a lot for it to know and understand us to be that person that can support us. Where's the knife edge on this? Yeah, it's actually a really, really good question.

I don't want to stand here saying that I've got the perfect answer. There are a few things to think about. For me, the framework that has helped me most when I think through these questions is to try and put this back to what we know. So it's not like we've had children in the physical world for a very, very long time.

And parents have started, and even as a society, I'd say there are norms that we have developed that we are okay with, where at some point you happy to let your kid play outside maybe in your field of view depending on where you live if they can go just outside and there no street like something like that And then at some point you start to be comfortable leaving them with somebody else and you might not see anything. You don't know what's happening. Now it's somebody else's responsibility to a degree.

At some point, you start to let them walk to school themselves, but you want to make sure that you know that they arrived. And like we have all of these things that we kind of build up. And I think what we need to get to is a clear set of guidelines of letting a parent or making it easy for a parent to be able to not to have to say yes or no entirely. So you're either treated like an adult or you're just not able to use it.

And it needs to be more of an on-run. We need to be able to give parents more tools where they can say, okay, right now, yeah, I'm going to let you use ChatGPT, but so you know, I can see every single conversation. Or even better, I mean, we have these AIs for a reason now. If it is a child account, every sensitive chat gets flagged.

But done in a transparent way. The child knows it's there. And then, and of course, at some point, it's going to then hit the point where the kid is, okay, well, I don't want you to be looking at my things anymore. And then providing a way to an environment where the child doesn't want to go and do it secretly does provide the parent a way to feel more comfortable.

you can imagine things like at some point when a kid demonstrates they maybe it's them maybe it's the conversation with the parent where they where you no longer get to see the actual conversations but it still gets flagged hey something something important happened yesterday i'm not going to tell you what but something important happened just like what would happen in actually just like in school or just like if your your child was with another another parent somewhere they there was a situation that unfolded, they solved it there and then.

But the parent might still go, it's like, hey, I've got it. It was fine. Nothing to worry about. But just if they're a little bit, there you go.

So it's about providing that unwrap. And I think that is so important because as society, I really think we're getting this wrong at the moment where we are too often. We are basically, we're creating this environment where kids are not actually exposed to reality in any way on. I mean, now we're talking about social media bans under 16, right?

And I mean, that just means that when they hit 16, if they were to somehow actually do this, which that's a whole other if, it means that the world is going to look like they're going to feel like, wait, I was in this cocoon thing and I had no idea. And suddenly everything will fall on them in one go. They have to figure out which school they go to. They have to figure out what they want to do in life.

And And then at the same time, they've got all this other stuff to deal with, which is, I think we should all be able to agree that it's not the best that you create this cliff where you protect from everything and suddenly you expose them to everything else. So yeah, long-winded answer to, I think we have these frameworks, but somehow we're not paying them. And I totally agree. And it reminded me actually of something.

I remember getting an email from, I think it was Google, to say that my son had turned 13. This was a couple of years ago. And now I didn't have to approve because he was 13, but that they would still inform me when he logged into new applications and platforms. And that sort of element of, OK, I can see that he's logged into Zoom.

Fine. So later on, I might say to him, who are you chatting to? Yeah. Did you talk?

And it's that open conversation. And I totally agree. I think it's those kinds of safeguards. I think, unfortunately, it again comes down to that digital fluency for people of do our parents, do people of a generation where they're having children that are at technology using age, understand themselves enough about technology to put those things in place?

Yes. And I mean, this is where the irony comes in. So, you know, the number one complaint we used to get at Super Awesome for we built all this parental consent mechanism. Right.

So whenever whenever there was something sensitive. So, for example, a very valid use case. We were working with Pokemon Go. Right.

Massive thing exploded. Everyone was using it. Pokemon Go relies on geolocation data. You need to know where you are.

The entire experience is based on geolocation. So very valid use case. If an under 13 wants to use this app, we're going to have to ask the parent, hey, this app uses geolocation data. Are you okay with that for your...

If they say yes, no problem. But number one complaint we would get is in very colorful wording, why aren't you letting my kid just use this? I don't want to have to give you my permission. It's like that.

So there's also this thing around parental responsibilities. Hey, we're trying to put in place all the right things. But there is definitely also a parental responsibility that is not being executed in many cases. And that brings you back to why some people want to try and just put this into the law, because parents are just not doing their job in some cases.

Yeah. And unfortunately, I mean, I literally had a conversation yesterday because it was announced in the Middle East, in the UAE, I should say. They actually brought a law out, and now they have done the social media. And interestingly, they're doing between 15 and 16, they're doing it an almost abridged year is their plan to allow that understanding and to kind of hybrid the social media experience.

And again, we'll see. We have national identification cards and things, so I think it'll be a lot easier to implement. But they were very much of the, well, I don't understand why the technology companies aren't doing more. Why aren't they the ones putting this in place?

We shouldn't have to ban the children. It should just be that they're doing their job. And having worked with companies such as Meta and People, actually, there's an awful lot they're doing. And one of their biggest problems is that parents just let their children do it anyway.

And I say that as a parent. And, you know, the amount of nags I've had and twisted arms and all the rest of it is just parental peer pressure can be really hard. But there's also that line of some parents who just don't want it to be part of their responsibility. And that's really hard to deal with.

And so it's, you know, safeguarding in so many different ways is a really tricky and quite turbulent way to have to do things. But the exposure to understanding that, I guess, is also part of that change, isn't it? That's incredible. And I love the passion for the education side of things that you're working on here.

I'd love to ask, where do you see things going? Is that even I feel like you used to say to people, where do you see yourself in five years? Where do you see yourself in a month? yeah it's i mean it's such a crazy time at the moment so i'm i'm a soft believer i'd say it's like i mean there's a whole bunch of different things here but the i am genuinely positive on how much ai is going to allow us to do i think it's going to be extremely hard so i think i think the next two three years might be really hard i think a lot of people are going to be losing their jobs and companies are going to have to, there's going to be a level of competition that many won't be used to simply because once somebody else starts doing things in one way, you either lose your company or you try to then adapt and then everyone is kind of going through the same things.

Now, on the other end, I genuinely think we're going to be getting to a society where we no longer soon in say five to ten years from now we no longer have to work to produce kind of economic value in in the sense of like the basic needs that that kind of we're basically going up a step in maslow's hierarchy of needs in a way yeah and i because the same thing is happening from a robotics perspective extremely quickly now where where the same models are getting put into ai we can now build uh sorry into robots we can now build real-time generation generating worlds which is accelerating all of that.

So like the next, yeah, next five or 10 years are going to see the biggest progress we've ever seen. And where I am at is trying to, I'm personally now trying to take a little bit of a step back. And I say that because when you start having agents doing a lot of the work for you, there's this concept called brain fry. I don't know if you heard of it, but the idea being that if you're constantly at the highest part of your intellectual capability, at least I cannot function like that for an entire day no and so it's exhausting yeah because my agents are doing all the grunt work that means that i only get to do the really intellectual bit and so the problem is by by 10 a.

m 11 a.m like i'm already overloaded i can no longer do that type of work so recently i started doing kind of 20 minute sessions where i play piano nice because it regenerates my ability to be able to then go back to the computer, give more instructions back to the agent. And so it's an interesting different balance. And so one of the things I'll be working on, you asked where kind of I want to be, is trying to figure out a new balance on how do I do work now that all of these agents are part of the way that I do that work.

Because it definitely can't continue in the way that I've always done it before. No, absolutely. And it makes a lot of sense. And I think that's a lot of what I've heard is this idea of it comes back to that being more human is people enjoying life more, having that time back to spend with family, go for a walk, exercise, pick up a hobby, learn a language, play an instrument, those sorts of things.

is and what a joy that something can give that to us those things that we you know some people wait until they retire to be able to do again actually no it's part of being a human being and living and enjoying life is yes working work is not everything and actually there are some really special moments that we can maybe draw out of that I love that yeah fantastic it was been absolutely wonderful speaking to you, Josh. Thank you so much for your time. If anybody wants to know any more about your company, we'll be putting everything into the show notes for them and sharing them.

And yeah, one to watch. We will be watching this space with all of your change and information over the next coming months and years. Thank you. Thank you very much for having me.

a fascinating conversation and one that challenges not only to use ai differently but to question some of the assumptions surrounding it joshua's comments about water and energy consumption are particularly important there are genuine environmental concerns connect with the rapid expansion of ai and data centers but dramatic statistics can quickly use their context as they're repeated. A bigger lesson that sustainability's conversation must be evidence-led, transparent and open to scrutiny, not just driven by a single headline or comparison.

I know that's definitely something that I've learned today. The same is true in education. Blanket bans will not stop young people using AI. They may simply push the use underground, where learners outsource their thinking without guidance, safeguarding or critical understandings that they need.

And perhaps a better question is not how do we stop people using AI, but how do we help them to use it to think more deeply? That means treating AI as a coach rather than answer machine, asking it to uncover our blind spots and ensuring that access to it, meaningful AI education does not become the foundation of a new skills divide. A huge thank you to Joshua for such a candid and thought-provoking discussion. You can find out more about Mindstone through the links in the show notes.

And as always, thank you for listening to the EdTech podcast. Please follow, share and join us again for another conversation exploring the future of education and technology.