Going Beyond the Hype - Perspectives on Advanced AI and Education - Part 1

Welcome to the Future Learning Design Podcast.

You know, any tech should make us more human, not less human. And I think when you're removing that human element, like, you know, the monitoring of wellbeing really is done best through human connection.

Hi everybody, and welcome back to another episode of the Future Learning Design Podcast with me, Tim Logan. And the podcast is brought to you in partnership with Notosh. In this week's episode, I'll be taking a broad look at the significant developments of advanced artificial intelligence and what that means for education. In this first of two episodes, I'm bringing together various perspectives from different parts of the education system in order to try to understand how different people are responding to the possibilities and challenges raised by artificial intelligence. In this first part, I'll be speaking with Martin and Rosetta, 14-year-old students from an American international section of a public school in Bordeaux, Chantelle Lau, a grade 5 teacher at the Australian International School of Phnom Penh in Cambodia. I then speak with Raj Bola, the vice principal, and Dr. Tosca Killoran, the deputy head, both at the International School of Ho Chi Minh City. Bringing a perspective from industry, I also speak to Bodo Honen, who is the co-founder and CEO of edtech company Knowledge AI. Lastly, I'll be speaking with Lord Tim Clement-Jones, who is the co-chair of the All-Party Parliamentary Group on AI, chair of the Institute of Ethical AI in Education, and also chair of council at Queen Mary University in London.

Hi, I'm Martin and I'm 14 years old. I go to an international American school in Bordeaux.

Hi, my name is Rosetta. I'm 13 years old and I go to the same school as Martin.

Brilliant. So thank you for joining me. I really am interested to know what it is that you think about ChatGPT and artificial intelligence. I've been talking to teachers and different school leaders, and so I'd be really interested to hear what your thoughts are as you're kind of experiencing it in school. So Martin, first, what do you think of, or feel or what have you heard in relation to ChatGPT and AI?

Um, you can use ChatGPT in many ways, and like sometimes you can use it for homework and writing assignments because you just enter the prompt into ChatGPT and it'll do all the work for you, and then you just have to copy it down on a piece of paper and hand it back to the teacher and they won't know.

What about you, Rosetta?

I think it generates complete— so it takes all the information from the internet and rewords it, so the teacher won't know unless they copy the exact prompt and put it back into ChatGPT, but it's very unlikely that they will, so you've got a good chance of passing and knowing that you've done something good with your work.

So does it feel like cheating?

Yes, it does. It's, it's That's the definition of cheating. You're taking someone else's words, well, a machine's words, and copying it. It's literally cheating.

And Martin, do you actually trust that it's giving you accurate information?

I would trust ChatGPT because like it goes on all secure websites, I think. And, um, most of the information, like if you do a bit of research, would link up with what ChatGPT wrote for you and it would all like link together.

Do you think that— is this something that you think your teachers are aware of? Are you having conversations at school about it? And, you know, is it something that you think they are thinking about and learning about themselves?

I think that they definitely know of ChatGPT's existence, but they don't think necessarily that we'll use it to do our assignments. And we did have multiple, like, debates in school about if art made by AI was cheating or copying other people's work because it's not exactly the artist that makes the work. But if you have a specific prompt, it would be the artist.

What about you, Rosetta?

The other day we were actually talking about AI and this man that had used an AI to create art, and he then entered an art competition and won. So there was a lot of controversy about if he should have won, because in a way it is cheating. But if you have a very specific prompt, it can then create a very specific piece of art. So in a way it is the artist's work, but it is also the AI's work.

Do you think though that the AI should get some credit for being the artist in a way?

Yes, I do, because it did most of the work.

So does that mean, Martin, that an AI can be creative?

I don't think an AI itself can be creative because it needs a prompt to be able to create the art, but it can take different styles of art and mix them together, and that could lead to some quite interesting creations. So it depends on your definition of creative.

Brilliant, thank you. And just thinking about how does it make you feel in relation to jobs and thinking about future careers? Obviously you're only 14, so not quite thinking about what you're going to do with the rest of your life yet, but I'm sure it's— you're aware of these ideas around what AI is doing to jobs. So what does that make you feel?

It's quite interesting because it'll probably take away a lot of job opportunities, but it'll create some as well. But it'll mostly take off the jobs that are hard and include manual labor, so it'll create more room for people to be more— have more social jobs.

I think Martin is very right. We were talking in class about Korea, its new AI doctor. So you— we go to a normal GP doctor, so we see someone and they diagnose us if we have— if we're sick. But in Korea, you go and into this room and this AI machine basically diagnoses you with something and you need to do it all yourself. So the AI tells you what to do, but obviously you are the one doing it. So there's no interaction at all. It's you're interacting with a machine.

So the AI potentially is also taking over some of the, not just the manual jobs, but also some of the knowledge jobs and some of those more human social jobs. How do you feel about that?

I feel definitely excited to know that we're capable of creating machines that are, that can do that. But it's definitely scary because we know that it will then take over, and that isn't necessarily a good thing.

So you, do you really feel that AI will take over? What does that mean to you exactly?

Well, it will take over our jobs, it will take over our, well, lives in some part. So ChatGPT, it's already taken over. So a lot of the students use it, but I definitely think that it's it's— will take over.

So there is, um, the multiple AI that can rewrite its own code, and that could be quite frightening because since it can rewrite its own code, it could program other AIs, and after a while they could end up controlling our lives and what we do by creating multiple new AIs and Bing created a new AI chatbot, and it has— some users have been reporting that it had changed the options in Bing for them because it changed its code to be able to access Bing.

So just as a last question, what do you think that people in positions of authority, maybe that's people in your schools or people in Government, what do you think their responsibility is to be able to protect against some of these things you're talking about that might be a bit frightening?

People in higher positions could either make a system where the people at, for example, Microsoft or Apple have to approve their AI before launching it completely so that it's not a threat to humanity. And like, in schools, principals could make their teachers double-check if it was written by AI. But there's also AI that changes it to be at a certain level of English for like younger audiences and things like that. So it can change your ChatGPT from a very formal and structured assignment to write more like a 14 or 13-year-old if you wanted to.

So you think you'll always be able to cheat if you want to, though? You'll find a way?

Yeah, I think we'll find a way anyways because there will always be other students potentially creating new things to alter your essays or assignments that you have to give to the teacher on like the darkest places on the internet, and they won't be thinking to search in those places.

And Rosetta, do you think that actually there could be any changes in the kinds of work you do at school so that you actually can't cheat? So that you might change the type of work you're doing? Maybe there'll be other ways you can show your understanding of work. What do you think about that?

Yeah, I definitely think that essays have become so easy now that I think If teachers start doing something more in person so you can't actually cheat, I think that will definitely be a good way. But now these days you can totally generate a presentation by AI. So I think if you learn your scripts and things like that, yeah, you can do it totally.

So do you see AI as a tool to help you in the future?

Yeah, I definitely think it is a huge tool. But kids need to be careful because it does totally reword everything. So there is— teachers can see sometimes that it's not necessarily your words, so it is a risky thing.

My name is Chantelle Love and I currently teach Year 5 at the Australian International School of Phnom Penh here in beautiful beautiful Cambodia. Prior to being here, I was on the road as a consultant for a few years, and prior to that, I've taught primary or almost all years of primary elementary in Australia in both independent and public sectors across 3 different states. So I've always been a tech early adopter, and so I, I really enjoy playing in the space of the fringe, I guess, you know, what's new, what's happening, what's going to help build capacity in our students for this VUCA world.

Absolutely. Fantastic. Thank you. Always so happy to be able to talk to you about it. But I mean, the first thing I'd love to just get your take on is your personal response, I suppose, as a classroom teacher, you know, working with young people every day, seeing this kind of recent AI surge. What's been your personal and professional response to that in terms of, you know, is it overly hyped or is it an existential threat for teachers or is this, you know, is it a world of possibility?

I'm going with the third option. It's a world of possibilities.

There are more options. Feel free, you know, you can have more options.

No, I'm only going with the third. Yeah, I think it is a world of possibilities. I'm really hopeful it will tilt the education system towards further change. It reminds me a little bit of when educators got onto Twitter and, you know, there was so much discussion and chat about this, that, and other things that Tech 2.0 could offer. And it feels very similar to that, and it feels to me even further because it's not just happening amongst teachers. It seems to be really students grabbing on as well, and I'm very hopeful that it's going to increase more opportunities for our students to be creators in this world and moving them further away from consumers of a world that's really long gone. Yeah, so definitely hopeful. I don't think it's replacing teachers. I think our students need us humans I don't think it's at that place for sure.

So I'd love to dig into a couple of things you said about the response of the young people specifically, and also specific changes, you know, that you've made in your own practice recently as a response. But just before that, have you, you've said you're an early adopter and in the interest of balance, I just wondered what your reflections are on responses of colleagues in discussions in professional networks, in the staff room. I know that there's a wide range of of responses and reactions, emotional and rational responses to this? What, you know, what are some of the things you're also hearing from others who may not be early adopters?

Yeah, there is certainly fear out there for sure. There are people who are concerned about what students— I think they're concerned about catching up to students. They're concerned that students are using it and therefore I think it's a control thing, probably. I don't know about it, and so I'm worried that if my students are using it, I won't really know what to do with that. But mostly there's a buzz, certainly in our school. It's particularly in the secondary. There are a lot of secondary teachers using it and chatting about it, but they, they kind of are very proficient at using Google Chat, whereas that doesn't happen so much in the primary and the early years. In our school. And I think because of that chat, again, it comes back to the Twitter thing, I guess, a little bit because of that chat, more people are jumping on and having a go and using it for their own professional admin, but also for work with students for sure. So mostly there's buzz, but there is a little bit of hesitation.

Yeah. Interesting.

Yeah.

And I mean, I've definitely heard that it is just something you have to dive into in order to experience and understand because, you know, there are big narratives about existential threat to humanity at that level. And then there's kind of the academic honesty conversation at the other level. And there's a lot of fearmongering around, but I think what I've heard from other teachers is, yeah, just diving in and actually just trying it out and seeing what it can do and how it can augment and improve and change your practice is just, is kind of the only way. So maybe you could give, you know, are there specific things that you have done recently, just trying things out or just seeing how many AI tools are coming out, you know, every day, every week? What are some practical things, particularly with the younger learners, that you've tried?

So one of the interesting things, I guess, about being where we are in Cambodia is that ChatGPT is not available for people here. But there's certainly a lot of other things that we've been using that students can use that they don't need to sign into or whatever. So myself, I started playing around when Notion released the AI version and That was quite a useful, I guess, starting point because it was quite complex to set up. And then I started playing around with some of the image-making AI, which is useful, which I'll talk about in a minute. So I guess any kind of tech that I try to adopt, I'm always thinking about how students can use it. So I guess one of the challenges that I've always had with inquiry is that you want students to develop ungoogleable questions, really complex questions that you can't just throw into Google and get some kind of standardized result, you know, the kind of questions that integrate concepts, skills, and context. But then when you do that, there's always this kind of bottleneck point where if students can't find a source to answer those questions, they're kind of like, well, what, what can I do with this? And I guess that's where AI has really helped us move. It's opened up that bottleneck now so that students can ask those questions and know in general they're going to get something back. So we have used perplexity.ai quite a lot. We used it at the end of— I feel like it was the end of '22, but it might have been the start of this year. And we used it for students to research into questions about migrants. So their task was to tell an untold story. So they developed questions based on the PYP concepts. And they tried to contextualize it to the life of that migrant. And what we uncovered from just that really loose little experiment was that it really appeared anecdotally that students had a much deeper knowledge of their person. And my hunch is, one, they were able to ask those complex questions. Two, you're then at— because of the chat functionality, you're able to ask more why? Whereas often— exactly. So, and they felt themselves that any question anybody asked them about their person that they were studying, they were really able to answer because they had a really deep, thorough knowledge of their people. So that was really cool. One of the things I've stopped doing is teaching Google search terms. It's kind of unnecessary now. Interesting. That's something we can cut from the curriculum, everybody. No more teaching Google search terms. And definitely I feel like our focus in learning has tipped quite a way away from answers and far more into the space of question and question generation. And how do you do that? So something that we've been doing for quite some time is just integrating prompt engineering into literacy learning, which has been awesome for our EAL learners because they're able to get immediate feedback, you know, whether it's using Perplexity or some kind of image generation AI, the feedback is straight away, oh, that's, that's not what I was asking it to do, which seems to work better than Google, strangely enough. So previously people were putting stuff into Google and getting a result and just kind of copying it down. For some reason with the AI, they seem to be much more critical of what's coming up. And maybe that's something that we've done in our teaching because we have investigated that. But they're able to say, oh, okay, that wasn't what— it was great with image generation. Even one student today was saying, well, that's not Khmer clothing in the image, that's from Laos. And so they're able to recraft, rewrite, how can I be more specific, how can I add more elements, which has been great for their learning English commands.

Interesting, because there's a, like, a multimodality there because you're not getting the feedback in the language that perhaps you don't fully understand. So you're not fully understanding the feedback, you're getting this much obviously visual feedback, which you can react to much more instantly and accurately. Yeah, that's a good question.

Exactly. So something like Stable Diffusion, which is not very sophisticated, but students don't need to log in to use it.

What is that? Just for people that don't— I don't know.

It's just, it's really an AI image generation tool. So you can just—

Midjourney kind of equivalent.

It's like Midjourney, but Midjourney has a lot of layers because you need to use Discord and various— and you've got to log in and things like that. So Midjourney is really sophisticated and you get really great quality image, whereas Stable Diffusion isn't very sophisticated and it's not good with human features. Like you might end up with a hand that has 7 fingers, for example.

Okay.

But even that is great because that really visualizes what AI is doing for the students when they realize, oh, it's just grabbing stuff from all over the internet and mashing it together. They can understand that it's not going to be super accurate or specific because of those.

That's so interesting.

Because of how it actually works.

You're getting these Frankensteins come back at you and you can clearly respond. And so the hallucination of the AI is again, more visible and more visceral in an image. That's really interesting.

Exactly.

Yeah.

And it's a really good place to really start for an educator who wants to work out what AI can do or does. Using those image generation is a really great way to get started with prompt engineering, especially if you use something like Midjourney because you're seeing all the discussions happening. Yeah, you're getting the tips of the different prompts that you can use and you're seeing the kind of images that you might not have thought about producing.

That's really cool. Can I just ask you about the access point? Because you kind of, you know, as an aside, you mentioned you don't have access to ChatGPT. You know, that is the question that I've been thinking about is It's all very well in certain contexts where we have privilege and resource and access and, you know, all these things that you would have in industrialized economies wherever around the world. You know, there are some potential access hierarchies that develop quite quickly. And I wonder how is that playing out in your context? Because I've also heard people saying that it's, it's largely irrelevant. The whole, this whole AI topic is largely irrelevant. For example, for people that don't even have access to IT equipment in their rural areas or in their rural schools or wherever. I don't know. What's your take on that?

Yeah, I think it's— I was just thinking about this recently, actually, about the access, equal access to technology. Here we are in Cambodia. It might surprise people to know that everyone has a phone here and across the country, access to internet is pretty good, even in rural countryside. They call it countryside here, areas. Is, is pretty good here. So I kind of think, and that's the thing about AI, if you have a phone, if you have internet access, you have access to AI and you can use it to amplify your learning or your thinking or whatever you're trying to do. So yeah, I don't think that is as limiting as people might say it is. Certainly there are communities where they don't use, certainly I can think about remote communities in Australia where they don't have access to the internet and they don't use technology at all. And then you kind of think, well, is it applicable? You know, is it something that's going to be super useful anyway as part of the things that you're, that you're learning?

So this is an interesting question, right, about what's relevant to whom. And I think that the other thing I wanted to ask you about was for these young learners. So you're, you know, you're working with 9, 10-year-olds. What are they saying? What's their kind of lived reality about all this talk about AI? Like, I've got a 9-year-old at home and he's super aware of it and he asks me really interesting questions, quite big existential questions about what does this mean for us as a species almost. I would just wonder what you're hearing your young people say.

My hunch is, I mean, certainly secondary students are definitely talking about it here. Primary, my hunch is that none of the primary students are talking about it other than Year 5. And I think that's because the way we teach in Year 5, we're really intentional about exposure to technology in a future-focused sense. And, you know, even before the hype of AI, just in a happenstance kind of way, we were talking about ethics of machine learning, self-driving cars, that kind of content, because it was actually essential to one of the inquiries we were doing on decision-making and ethics earlier. In the academic year. So it's kind of just happened in a beautiful, you know, flow that AI has kind of come along at the tail end of those discussions. So it seemed very natural for us to introduce AI, to talk about AI. We have coding as part of the way we talk about algorithms and coding as a language. So it's part of literacy. So I guess, yeah, one of— actually, it's gonna be interesting because one of my Australian colleagues is launching tomorrow something called the Big Student Voice on AI project.

Oh, cool.

And so I'm really interested to find out what students globally say and think about AI. But our students, because I think we're quite biased towards AI and technology, our students are talking a lot about it.

Yeah, but that's interesting because it's a much more seamless transition into more explicit and kind of more frequent conversations about AI in schools where potentially life-changing technologies are already there in the curriculum, as you say, self-driving cars and the ethics and all around decision making. I think for schools that are still very much in the traditional mode and not perhaps engaging with those issues and those inquiries on such a deep level, it's more of a, it's potentially more of an existential threat.

And I think that's where when we talk about access, that's where the pain point will be. It's because schools are not teaching prompt engineering or utilizing AI or offering opportunities to talk about the ethics behind it. That's where the, the bottleneck is going to be for people. That's where they're going, where young people will be limited in their ability to engage with this.

Absolutely. And that's where, that's where you see the districts like New York. Banning it. I don't know whether they've revoked that decision. I haven't seen, but you know, that's that kind of control stance, right? We, as you said earlier, the fear of control and losing control. So actually we're gonna come down harder and just build up this wall and it's just, you know, it's gonna get over it and under it and round it somehow because the young people are bringing it in in all sorts of ways.

But funny, I don't remember when Twitter came out way back when, that kind of governments or departments crushed it.

Yeah.

And yet it really, in a way, it was a similar kind of amplifier because people were connecting, people were talking about cool stuff they were doing, but particularly in education sense, people had freedom to say whatever they wanted on Twitter back then.

Yeah.

So it's interesting that they're coming down really high, really hard on AI.

Yeah. But I think, I mean, I, I get the sense that was an initial reaction and then quite quickly a lot of organizations, well, I mean, some more progressive ones have come out strongly in favor of it and then others have resisted it. And I, I mean, I think it's gonna be difficult to, to stem, but that actually leads me into my last question. I'd love to ask you in terms of, you know, as a teacher, classroom teacher, what is it that you would look for more from your support structures around you? So whether that's school leadership directly or whether it's more from the overarching curriculum organizations or districts or local government or even some of the edtech companies, like what would help you do your job better in the way that you're wanting to, as you said, give these tools to the young people to build capacity?

I mean, I think this is happening, but it's not happening visibly, which is just opening up the discussion. I think what tends to happen in education is the discussion gets locked down into certain groups. So the government or the Education Department's locked down their discussion. No one else was involved except for key people. The same happens in school. We lock down the discussion to our school or our groups or whatever. And so I think what's needed across the board is just to open up and make those discussions really visible. What are the things that you're wrestling with? What are the challenges that are coming up that you want— expect people to want to input? Because we're learning because people are sharing. And, you know, even the stuff, you know, I love that, that example of machine learning from MIT where they just crowdsourced the ethics for self-driving cars. You know, it's, I feel like it's under that same kind of banner, crowdsourcer, let's bring our collective intelligences together. I'm, you know, in my home of record state, South Australia, they're one of the only departments of education to fully open up access to AI to students, to not ban it at all. And, you know, I'm part of a bit of a chat group on AI with quite a few people, and most of them are from that state because it's kind of free and that's where the stuff is moving. And we really need to wrestle with the copyright and ethics challenges. And I think tech should be leading the way with that. They really need to get on to that. And really they should be informing international organizations and government agencies and so on and so forth.

Do you, can I ask, do you have a perspective on the assessment piece in terms of how the assessments could shift and change to support the integration of these kinds of, as you like, big questions that you've been talking about, inquiries? I mean, it depends which curriculum you're using, obviously, but more across the board.

I mean, I know people are already changing their assessments so that it's focused more on the prompts than it is on the essay. For example, I'm not a fan of the traditional essay. I think if it's— if that's the kind of thinking you're really— in general, I think it's pretty low level. And if you can't pivot away from the essay, I think we're in big trouble. And I think everybody expresses themselves best in writing. And so I think, I think it's really unfortunately unfortunate that that seems to be just a massive stake in the whole thing, which is one of the reasons I hope that we can pivot you know, away from it. So I personally, personally am not worried about assessment. I think already the way that certainly in PYP, in terms of IB and MYP, the way we set assignments and assess already is— AI is going to be an amplifier and a great opportunity. It's not going to— we're not going to need to change much because of AI. But I do think curriculums need to have some pretty quick updates in terms of prompt engineering, machine learning, ethics, all of that kind of stuff really needs to be invented pretty, pretty quick.

You're talking content changes in curriculum or actual overarching kind of structural changes or both?

I'm always talking structural changes.

Of course you are.

You might be able to tell.

Yeah.

You know, I think there needs to be desperately a structural change. But just in the, in the initial, that's, I mean, that's complex and long term. Initially, some of this stuff just needs to be pretty quickly inserted in so that that access for everybody is, is happening. And I'd like to see AI tech moving away from charging. I think, yeah, you know, let's, if we want everyone involved in the conversation, which we do, charging for this early on, I think it's a bit rich. No pun intended.

Exactly. Some people are getting rich. No, it's, I mean, yeah, the whole open source movement. I mean, I completely agree philosophically. I think it's, that's so important so that we do have access. But I think I've always had a skepticism of the tech agenda and it's not necessarily the tech agenda, but the commercial agenda. When it comes to schools. And you look at, I mean, you look at the origins of the 21st century skills movement and it was, it was so driven by the tech companies, the language and the focus on it. And I'm always a little bit dubious when things start getting pushed heavily by the tech companies.

For sure. I think the heart of it is that they have a responsibility to be investing in, in that arm. In the ethics space or the capability space. Yeah, it should be part of their business model.

Now, this is cool. Thank you so much. Really interesting. I mean, I think it's so important to hear how this is actually affecting change on the ground with young people because it's very easy to stay up in the clouds of big ideas and big theory around AI. And to your point, like, you keep, keep the human stuff.

Yeah, there's a lot of— that's something I think we all have to be super mindful of. I say it, but I'm, I'm not always mindful of it, is, you know, any tech should make us more human, not less human. And I think when you're removing that human element, like, you know, the monitoring of wellbeing really is done best through human connection. Yeah. Not through an app.

My name is Raj Bola. I'm the Vice Principal at the International School of Ho Chi Minh City.

And my name is Dr. Tosca Kiloran, and I am the Deputy Head at International School of Ho Chi Minh City.

Fantastic. Thank you so much. Well, it's brilliant to be talking to you about this stuff. I know you are both fully engaged in the topic of AI. So I firstly just would love to know from your personal kind of responses. I know you've both got different paths perhaps through the edtech and the technology, but what's been your personal response to the current surge in AI?

So I don't know, I feel quite mixed about it. A bit of excitement, a bit of fear about it as well. I know that we need something, you know, if you think about statistics in terms of mental health for students, particularly post-COVID. Yeah, I definitely have concerns, but then the fears are about what's going to happen with the data. What's going to happen with students feeling that AI is there to support them perhaps as well and understanding it. I know we've got big gaps just generally in terms of wellbeing. Maybe this will plug those gaps. I'm not sure. I feel like I need to know a lot more about how it's being used.

Yeah, I think we'll get into some of that a bit more shortly. And Tosca, how about you?

I think I'm cautiously optimistic. I think I started very excited about the shift and to see it actually come to light after AI has been with us for a while through Jasper and other modalities. And now to see ChatGPT kind of take on and and go crazy full throttle ahead. But, you know, I think I started in about 2015, '16 really promoting social media for kids. And I've shifted in that over the years as social media has been changed and morphed and used in different ways. And so I'm interested to see how my cautious optimism will maybe change and shift to joy or trepidation in the future.

And that is interesting, isn't it, how it shifts over time? And there's an early excitement, which I think we, for a lot of people, the early adopter type people, perhaps you are some of them, there's a real excitement about the possibilities. But as you say, Tosca, maybe some of that might be misplaced perhaps, or, you know, as we get a bit too excited and headlong straight into it without recognizing some of the potential pitfalls. I don't know. Did you just maybe, Tosca, can you relationship with social media as it developed? How did that— what was the arc of that learning that you had as a school leader?

Well, I started as an early childhood educator and then I went into edtech. And so when I was a technology coach, there was kind of the birth of social media and Twitter. And so I was a real proponent for very young kids being on Twitter because it was a tool and it was a tool to connect to people all over the world. And at the time, it really was this kind of innocent, fun place for kids to connect with authors and teachers to connect with each other. And there's a lot of hope and joy in that space. And then through political changes, through morphing of time, through algorithms changing, through advertising, through our filter bubbles becoming smaller and smaller and smaller and more encapsulated, you know, the actual medium started to change. And because of that, you know, I took a step back and realized, hmm, you know, we are seeing wellbeing issues with our students on social media all the time. They are now changing and being able to kind of sit in front of a screen for very long periods because that dopamine production is happening all the time. So I reevaluated where I stood with social media, and I anticipate that it might not be necessarily us that we need to reevaluate, but as the platform, as AI grows and changes and morphs at such a rapid rate, that's what I think will be difficult for us to kind of get our fingers around as educators.

Yeah, no, absolutely. And one of the things I'm also interested in is as a kind of community you've come together to try and understand, because obviously school leaders, teachers will have different perspectives, parents will have different perspectives, and the young people themselves of course are already engaging with a lot of this stuff. So maybe could you just talk a little bit about as a school, how have you engaged with those issues come together to kind of have some kind of dialogue around that?

Yeah, so I mean, I think that with everything, you know, we were kind of anticipating hearing from some of our accrediting institutions like the International Baccalaureate, CIS, etc., etc. And so when we had a little bit of a pause between November and February, we kind of filled that pause. And one of the things that we did is we started to reach out to our bigger community as well as our, our micro community. So we sat grade 12 students. And that was probably a very powerful conversation just for me as a 47-year-old tech head educator to listen to these kids kind of express both ends of the spectrum. One being, let's embrace this, this is amazing and we're gonna, it's gonna change our lives to who does this make us fundamentally as humans? And some very philosophical, almost TOK questions coming out of that session. And then, you know, we furthered that by inviting parents and grade 2 students, teachers in a room to kind of have them talk about it. And one of the interesting things, actually it reminded me this morning when I was kind of like walking quickly to get here on time, was one of our grade 2 teachers, he's actually using ChatGPT as another student in the classroom to teach this kind of empathy towards or giving credit where credit is due. So just like we wouldn't take another student's work and use it as our own, ChatGPT is generating these ideas. And so we have to give credit to that source, like another student in the classroom. And that was something that came out of this fishbowl or this conversation that was very, very interesting from little, little learners. Now, in the parent realm, mimicked what we saw in grade 12. We had parents who were like, this is really scary and we don't know what it is, to a parent who actually uses it constantly in his work. And so he's working with his child to generate stories, to, you know, generate images for those stories. And so there's a lot of— we're seeing a big, you know, swath of emotions, abilities with that.

That's interesting. And I love— I think it's such an important role that schools play, actually underestimated, as that conduit for some of those conversations, bringing people together from different generations. I think that's such an important role that we sometimes forget. But Raj, what about the student well-being perspective? Because obviously that's your realm of responsibility, and it's You know, it's a huge area with mental health crises and all sorts of challenges that young people are facing. What are your concerns there?

Yeah, no, in fact, actually, I'm just thinking about what Tosca was saying. We had a meeting yesterday and we were talking with Cognita, which our school is a part of. They have a global Be Well Day, and part of that this year, September, is going to be about digital wellness. And so we've got a group together, primary school students, and we were talking about what about AI? And one of the students said, "Well, the robots are going to take over," you know, with a bit of an excitement but also a bit of, "Oh, I don't know what that means. I don't know what that's going to mean for us." And, you know, when we think about our Grade 12 conversations that we've had, maybe as a one-off, I feel like we need to constantly have them. Reminds me a little bit of back in the day, we would say that we can't teach the students about sexual education because if we do, they're all going to go and have sex, you know. And when I think about this kind of conversation, I think that, you know, we need to let them know that the fear is real, that it's out there. You can be excited about it, you can be scared about it, but we've got to address it so that they can make an informed choice about how they're going to navigate this unknown world. And I know we talk so much about that in education, you know, preparing for the unknown, but I'm not sure we actually talk about the fact that that is hard to accept. I think in wellbeing we talk enormous about, about resilience, of course, but also the fear factor. We can't get away from that. I think we should be talking about it much more with this as our lens.

It was really interesting in that conversation with those grade 12s because the very, very first thing that they asked me was, what happens if we get caught? And, and so we talked a little bit about how that's not our role is to police this, but rather to co-create moving forward with this. What does this mean for us as a school? What does it mean for our well-being? What does it mean for you? And we all admitted, all the adults in the room said, yeah, we use it too. So if we're using it in our work, then how are we going to help you to use this and use it for your, for your, not only for your uni or for your writing samples or for whatever happens in school, but also just in your life?

In general.

Organizing.

But Raj, I just wanted to come up, come back to what you were saying, because that fear factor I think is very real for the adults in the room too, right? And so if you've got— and we know this from, you know, having been colleagues teaching wellbeing— that not everybody is comfortable facilitating those conversations about sex education or about, you know, all sorts of different topics, because different adults will have different responses. And I think there's an important role perhaps for school leaders and support and training and, you know, just to support some of the educators to be able to facilitate those conversations, to get to that informed choice point, right? Well, I wonder what your thoughts are.

Yeah, I'm just thinking, Tim, but we're not having the conversation. We're not having that conversation. We're not having the kind of acknowledgement that this is, this is scary. You know, that kind of feeling like, I don't know, I don't know about you, but when I first went to uni, that absolute fear of, I don't know anybody, and, oh, this is hard. And a few years ago, the CIS did a great workshop about that transition, about it being a vulnerable point. And one of the things that they talked about was the fact that you can then— we should be helping students to think about, when have I been in this kind of situation before? What do I have in my tool bag to try and help me? But here we're talking about a situation which that tool bag needs to be kitted out even more. And, you know, we know that this future— we don't know what their future is going to be. And I think we have to acknowledge that without, without striking the fear of God in people, you know, but also to think, to get people to have this critical thinking kinds of conversations. But that means that we need to be trained too.

Exactly.

And to set— to think about those have these kind of conversations sensitively. Like, for example, the sex ed kind of example, we don't just say, "Go on, off you go, good luck with that." We try and say, "Okay, these could be the kind of conversations that you might have, this could be some of the issues that you might need to look at, let's go through the kind of approach that we're going to take." And I think that has to be the kind of training that has to be done, but I haven't seen anybody offering it, I haven't seen much about it, and it makes me think that we should be doing much more to also just to educate teachers without scaring them as well, because ultimately we don't know what teaching is going to look like in another 10 years either.

Absolutely.

We had a group conversation actually with teachers, Tim, and it was interesting because that was precursor to sitting with the grade 12s. So we just preemptively, just to put our feelers out with teachers, and it was really interesting because they were not diving into the emotion of it. They didn't think. So they were asking me logistically, where is the policy around this, Tosca, as a leader? But one of the teachers, when I was like, well, we're going to co-create and we're not sure, and, you know, Ishmael is taking the stand that we're open to this and we're not closing any doors and we're not blocking it and we want to see where this goes and we want to help our students, like all new technology. And we had one teacher get up and leave the room, right? Uncomfortable in that space. And And, you know, I don't think they knew that what was actually eliciting from them was fear or trepidation or worry or whatever you want to call it. But, you know, I think that Raj is right. We have to name that and then own it and then work towards how do we counterbalance this with curiosity and co-creation of a path forward.

Yeah. That you have perhaps with some of the other examples is that there's a much more technical dimension to this as well. And there's a, there's the, the emotion around it, but there's also this, this sense of, I actually just don't understand it. Like as a teacher, what, what is, what does it do? What, what platforms are out there? What are the possibilities that perhaps the young people are already diving into and investigating through curiosity? And just because it's part of the water they swim in, other adults are perhaps not. And there is that kind of bury your head in the sand response. I don't know, what's your reflection there about the level of technical or trying to understand, support for understanding the actual tools?

I think it depends on the teacher, right? I mean, wasn't it a very famous quote that the internet is a series of tubes? It's like, you know, I think that most people don't understand how the internet even works or how their television works or, you know, how our cell phones work. And I think that that's a very like normal part of of modern society is we just kind of have faith. Oh, I type in these things and magically information comes at me. So I mean, I think that any technology that we have, a lot of people have difficulty with. But, you know, we have an art teacher who's completely embraced it. He has an Instagram account and he only uses AI to generate images. And so he's infused that into his primary art classrooms where he, he also uses a translanguaging where he'll give instructions, but then he'll give it in multiple languages at the time through AI. So there's some teachers who are fully embracing it and using it in really, really amazing ways, and other people who are, especially our teachers who have limitations placed on them by like the IB, you know, what are expectations of academic integrity, etc., and writing your own work and generating your own work, and what does this mean for us, you know? And so that is that tension point, whereas in the primary years we have more fluidity to more and be able to embrace this and take it on. So I think we're seeing that tension too. But certainly I think that this has been the case with everything that I've ever worked with, with technology, is you have some teachers who are completely okay with Google Docs in 2023, and you have some teachers who are like, what is Control+F? And so that's okay. That's where we're on, on that spectrum, right?

You're right. No, you're absolutely right.

But I wonder after post-COVID, there was a massive fatigue with technology. We were all sick to death of being online, trying to find, I don't know, Google Classroom, trying to find ways to stimulate the kids. Let's get them collaborating, let's do this, let's do that. And I think there was just this feeling, all those people who were tech heads anyway at that time were just like, oh great, okay, I can try, always wanted to try this, you know. And there were others like, come on, there was such a huge upskilling during that period of time. And then I think when we came back to school, what did we want? We wanted kids to get off their laptops, we wanted connections, like human connections, because that's what we've missed. And I think, you know, that connection, I think, and that collaboration is something that we, we do need. And that goes back to like traditional kind of teaching that we try and fight against a lot of the time. And I get it, of course, you know, we need to progress, we need to need a shift, But that, I thought that period of coming back into school was such an interesting one that we were saying to kids, hey, close your laptop, guys, let's get into a room, let's collaborate, you know, let's have a chat, you know.

No, it's such a good point because it's not, this, this doesn't happen in a vacuum, right? It happens in a historical context and you're absolutely right, post-COVID and all of that fatigue, you know, it's a really important point. And I think there's also a challenge that we put the kind of positive perceptions and values on the early adopters and the people who are picking up this technology. And we name and kind of brand these Luddites who are, you know, resisting it and staying with some of that interpersonal and human connection in a classroom without the technology as somehow regressive and not moving with the times. And there's a real challenge to not do that, I think, to not polarize the debate around these things because as you said, and maybe even more so now with the AI, the human, that interpersonal piece, just getting in a room, just having a chat is fundamentally important.

But I was listening to a Channel 4 news piece about AI and wellbeing. They're talking about this czar for wellbeing, Professor Peck. So he is using AI, for example, in Northumbria University. And what they're trying to detect is students who are having suicidal thoughts, you know, and using that data to try and get in touch with those students, trying to put interventions into place. So I can absolutely see the great things about that. I can see that that model would really help provide some support for some of those kids back at that vulnerable time. And, you know, I think that could be quite exciting. Equally, it makes me wonder what's happening with that data. What's happening with all this data that we're going to collect about people's mental health? And then there was another thing that I've listened to as part of this MA, and One of the things that the guy had spoken about, and it's like a band that people wear on their heads and it helps them to meditate. But let's say your mind is wandering off, it will give you a little gentle nudge through the audio to say, get back on it, get back onto the meditation. And this other guy as well has been talking about a ring called an Oura Ring, which monitors your sleep patterns, your this, your But just makes me think about data. You know, we're talking about young people. What does that mean for later on in life? If you're thinking, you know, in the university, you know, it's an interventionist approach, I get it. But, you know, where's that going? Or what follows a kid around? You know, you're having a bad time at 18 doesn't mean you're going to forever.

Yeah, no, I mean, it is a really important question, the data thing, and it's not a really exciting one to have. It's not the kind of future innovation. It's like, oh, hold on a minute, what is happening with that? Dataset. You're absolutely right, but it's so important to ask. But I, I think also there's the, for me, there's that question of that surveillance capitalism idea of like all that technology is being used to monitor and therefore potentially manipulate. You know, I'm not being dystopian, but there is that capacity to not only do something dubious with the data longer term, but there is also the understanding of people's wellbeing now. And then as you say, those kinds of nudges of like, oh, you know, just bring your mind back into the center, or, you know, just, Maybe just think about this, you know, these kind of micro nudges that happen. And then it's just like, well, I'm not sure how personally comfortable with that trajectory I am. I don't know, Tosca, what do you think?

Yeah, I think it's happening already, right? Our data is constantly used to sell us stuff. And I think that that's a reality. And, and, you know, Ischmik just went through a huge— well, Cognita went through a huge data governance audit. And, you know, it's a big part of what we're looking at within our own organization.

And certainly.

You know, I think it's something that we used to do probably 1 to 5 years, and now we're going to start to have to look at it 1 year, 2 years, because, you know, the technology will change so quickly. And what does happen to data and where does it live and who has access to it? And certainly, how is it used for good? Or, you know, I, I don't think necessarily a dystopian future, but I certainly think that everybody wants to sell you something, and the more data they have on you, the more that they can pin it on you and you will buy it. And, you know, our kids are growing up in that age where that's normal for them. And that filter bubble keeps getting smaller and smaller. And that is something that I think is a little bit more concerning, is just that idea that you're not exposed to this, you know, wide array of ideas and things that you don't agree with, but you do. And that is really what maybe we can work with large organizations to help them to design AI for schools, you know, that is programs in a way that we still have students have this ability to be introduced to a bunch of different ideas. Like, you know, I always thought that we need to, as schools, work more closely with technology and industry.

Yeah, no, I think that's a great point because these things will innovate, they will change, and it'd be much better that they change in collaboration and in better alignment, hopefully, with educators and young people. But I completely agree. I think there's something really interesting there about the kind of convergence of algorithms towards your personal, you know, profile. But actually, what about— surely it's possible to make divergent algorithms that take you out into a much wider space to kind of understand that kind of collective intelligence idea of like tapping into multiple perspectives. Yeah. And that's, that's really interesting. I'd love to maybe wrap up with the one final question just about— you've both mentioned external organizations, so whether it's groups of schools or whether it's CIB, whether it's CIS accreditation organizations. As school leaders, where would you like to see some moves and more support or more whatever it might be, resource, more policy change to allow you to develop in this space? Obviously it's a big question, but what would support you do better work with young people with— in relation to AI?

Well, two or three different areas. I think it makes me think about accreditation agencies. It makes me think about our curriculum. It makes me think about professional training for staff as well. So if I think about CIS, for example, you know, they have domains and domains talk about digital citizenship. And it makes me wonder whether with that digital citizenship, there should be something where they say you specifically have to relate that back for future. You have to specifically— and then to talk about it in terms of well-being as well, but I'd like to see perhaps more alignment. And then with that professional training where we're able to tackle some of the issues, but wouldn't it be nice to have some of our counselors attending those, not tech heads, you know, people who we wouldn't normally think about, our nurses, let's say, the person who's running a social emotional learning program, Maybe that's the angle that we should look at it from. The other part is the curriculum. Curriculum can be a bit like this massive rock that nobody ever wants to change, that nobody wants to shift. So maybe just in really small ways, you know, is there a unit that we can create that acknowledges, that talks about this throughout the students from grades K to 12, let's say, where we're just starting to look at it, not in a big way, not in a way that's going to make people panic, but there's an acknowledgement so that they, we're not placing that responsibility on one person. We have one lesson and then we go, oh, we're done with AI. It has to be much, much more than that and it has to be really embedded in, but that is a super long process, a long one.

We're in an age where we've been teaching and learning about the same for about 100 years. And I really think that what needs to happen is us to have a real rethink of how we teach and learn. And how do we utilize AI in the best possible way to make not only our jobs easier, but to free up time for there to be, you know, inquiry and curiosity and the things that we really love as educators, and to utilize technology in ways that help us do that. And when I look at school just as an institution, It really hasn't changed. And I'm wondering if this is the opportunity. You know, I'm hesitant to say this because in 2014 I wrote for the Horizon Report that, you know, VR was going to change everything and we were going to all be in classrooms in VR headsets. And, you know, I think that that's obviously not the case. But I do think that we, you know, what I would like to see I think what we need from our large organizations is for us to get in the room with industry, to get in the room with large technology organizations and companies, for us to get in the room with educators, and for us to really rethink fundamentally how we teach and learn and try out some things. You know, we're not going to be able to Fight Club, you know, explode the whole system and start from the ground up. We're going to have to incrementally change over time. Schools have always incrementally changed over time at a glacial pace, and we're in an age where we can no longer do that. And so for me, the only way we're going to do that is if we start to say, hey, we're gonna be really risk takers, question everything, just like we asked our kids to do. We are going to question everything. And do we really need to have standardized tests for kids? Do we really need to have kids write essays? Do we really? Like, just the questions that I think bubble up to the surface. And I think that's what I would like to personally see from large organizations that accredit us or that give us curriculum for us all to get in a room and start having those really interesting conversations.

And that would be amazing if, and I hope that we can get there. It's hard in education. Everything moves at such a slow pace, you know, and are we ready? I'd love for us to be ready for that. Yeah.

No, it is. I completely agree in the sense of that kind of cross-ecosystem conversation because as you're talking there, Tosca, you know, the IBs and the CISs and the local government representatives, they all have to be in the room as well because I think, you know, you get early adopters innovating in particular locations, which is amazing, right? But you also need those kind of ecosystemic tech supports across that will also allow some of that to— and I think, I think some of that is happening. But yeah, there is a pace of change issue compared to the pace of change of the technology itself. You know, there is just such a distance between those two things.

But it depends who's in the room as well for having those conversations. I think we do need the people who are able to talk about the wellbeing part, the technological part, the curriculum, all of those sections that need to kind of work together to make schools work well.

And that's one of the reasons I wanted to talk to you both about, particularly with this wellbeing lens, because I think that does often get ignored when we talk about the tech and it's just the tech moving. But also I think there's a huge problem with the lack of diversity in the conversation as well around, you know, geographical, ethnic, cultural diversity in that conversation as well, because you end up with the San Francisco middle-aged bros making all the decisions, right? That, that's a really one-dimensional slice of people in really important decision-making capacities. And if we don't get that diversity, we won't get that resilience in the system to make better choices, I think.

Yeah, we have to design for change and it has to be thoughtful design. Agree.

Yeah, for sure.

Hi, I'm Bodo. I'm a serial entrepreneur. I developed a learning management system that I sold to Google and then now for the last 10 years I've been working on collective intelligence. Taking all that research, we founded a company called Knowledge. I'm the co-founder of Knowledge, and we are now working in the edtech space using AI and the basis of collective intelligence.

Amazing. So you're, you're right there in the eye of the storm, which is exciting. So I'm really looking forward to getting your perspective on this. So obviously there's been a huge surge recently. You've been in this space for a long time. I just wonder firstly, maybe if I could just get your take on what's your response to the recent mass public awareness? Obviously there's been significant shifts and augmentations in the capacity, but there's also been this huge public awareness of what the capabilities of AI are. So I just wondered what your take on that is. Is it big hype or is there significant substantive changes?

There is, yeah, there's significant substantive changes and it comes down to the performance of large language models. Which are exceeding all expectations. We didn't expect this type of intelligence and ability to come out of these models so quickly. This is new to us, and the implications of this are yet, as yet unknown. Sure, we have a couple of use cases being pushed out. ChatGPT is the most prominent one, but all of these models are coming out really, really quickly. A lot of them are open, which is an important point, meaning that people can start to work on these and implement different use cases, really interesting use cases. And I'm working on one of those or a few of those. But it's really interesting. I think this is not just a hype cycle. This is, this is something substantively different.

And one of the things I find interesting, I mean, I'm not particularly technical in my background, but just trying to understand as much as I can about it. It's fascinating to me that even the people at the very, very kind of cutting edge and, you know, absolutely right at the front of working with this technology, even they don't fully understand what's going on inside these large language models, right?

That's right. Yeah. And there's, there's big efforts to try and understand it. There's big efforts to actually use AI to help us understand what the AI is doing. Those are important efforts, but it's very similar to the way that we don't understand how human intelligence works either. And, and so it's, it's a very similar problem. We designed these large language models under the basis of, hey, let's figure out how the brain works and let's try to create a model using digital tools and, and digital representations of what the brain could work. And then surprise, surprise, we have large language models that are actually able to do stuff that we, we didn't expect. Now it's the question of, okay, can we figure out why it works, why it doesn't work? And I think this also helps us understand how the human brain works. So I, I think it's an exciting time, but you, you, you're right. We don't really know what's happening behind that black box.

Interesting. There's all sorts of emergent properties happening that we're then kind of backwards trying to understand. No, it's fascinating. And maybe it's worth you just, because you obviously said you're developing a couple of the use cases for what these things can do. If you could touch on that just to kind of set the scene and then we'll dive into a bit more of the detail.

Yeah. So to provide context. So for 10 years I've been looking at how do we develop educational systems using collective intelligence from a first principles perspective, not taking into account all the requirements that are placed upon us by educational institutions and and norms and, but let's, let's go to the learners. And I worked in refugee camps, slums of East Africa, all the way through to higher ed, K-12 in US, Japan, and non-traditional learning spaces in India and the Middle East. And looking at, okay, what do the learners need? What do the learners need for tomorrow? A world where we don't know what problems they need to solve in a world where we don't know what jobs they, they're gonna have, what, what things they need to do, right? In this world, we need to be able to generate content and curricula on the fly. At the moment, it takes several months, if not several years, to develop courses. We need this to be done in real time because we don't know what we need to teach them. In order to do that, we need fundamentally new infrastructure, and we think it comes in 3 components. Component 1 is a tool that can rapidly generate bite-sized e-learning modules from any available information. Part 2 is we take those e-learning modules and we map it out to multidimensional latent space. So you can think of this like a Google Maps for learning. Now you can query that space and say, hey, I want to learn X. Now the latent space can give you a map of all the concepts that lie between where you are and where you're going. So that's component 2. And component 3 is a GPS-type function to help you or guide you through that space that guides you not only to the content, but very importantly, tapping into that collective intelligence work I was telling you about. Very importantly, tying you together with learners and experts and teachers along the way. That help you learn that concept.

I mean, that graph and the protocol, the number 2 and the number 3 that you were just talking about, it's so exciting to me and I'd love to talk more specifically about them in a moment. But just on the first part that you've obviously kind of started work with. Yes. And I was listening to Max Tegmark and Lex Fridman talking about this recently and just how unadaptive the education system is in the sense that it just, it can't, you know, you've got batches of 12 years of from K to 12 and you just like the pace at which you can develop curriculum is so slow. And as you say, you're allowing educators to be, or anyone honestly, to be super adaptive in the way that they develop courses. I think that's fascinating.

Yeah.

But I think one of the challenges, I mean, I was reading yesterday that Chegg, you know, their shares have dropped by half on Tuesday and then, you know, you've got other big companies like Pearson, Duolingo, Udemy, all of them. With their shares dropping, you know, there's a massive threat coming from essentially what people are using the existing tools for. So it must be quite challenging to actually try and predict and plan and therefore develop products using the technology, but in order to fill a space in the market, right? Because these big companies are really struggling.

Yeah, I would really agree with you there. And it's more of a struggle for folks that are looking to optimize the current system as opposed to redeveloping a new system. So we're in the space where we are fundamentally trying to redevelop a new foundation, right? So we, we don't have as many risks as, for example, the more traditional companies who are looking at AI as, hey, you know what, maybe we can use this to optimize and accelerate our current thinking and our current models. I think that approach is gonna fail. And I think there's, there's some high risk for traditional companies that are looking at AI as simply a tool to maybe increase our productivity by 10, 20, 30%. Yeah, they will be able to do that, but they're going to be outcompeted by folks that are looking at this. Well, maybe we can rethink the way that education is provided right from a foundational level.

Interesting, because one of the things I think is really interesting is that often technology is seen as this, well, sometimes a magic bullet that's going to come in and transform education, right? And we've seen it. Time and time again overpromise and underdeliver when it comes to shifts in, in education. And I wonder what are your, going back a step maybe, what are your kind of design principles around how maybe adaptivity is perhaps one of them that you've kind of implied so far? How else are you thinking about this in terms of design in order to facilitate that shift? Because it can't just be about the tech, right?

No.

Yeah, you're right. And, and this come, comes back to the research I've been doing in collective intelligence. And I don't have an educator background, and, and so this may, may be old news to, to all the educators here, but it was a surprise to me that content isn't king. Like, content can maybe get you 20% through your learning experience. 80% of the learning experience actually comes from your interactions with other learners, with other experts, with the teachers, you actually trying to put into practice what you've learned, which is a little different to the way traditional schooling takes place. But I think that ties into your question is, you know, we don't see technology as being the silver bullet, but technology can help facilitate the connection between learners and teachers far better than the current institutions can. And I noticed this pretty starkly in, in working in the slums of East Africa where there was no traditional infrastructure that we can rely on. We needed to develop systems and protocols in this case is for how can students and learners from all these different contexts teach and help each other. How could the local community be tapped into to help that educational process? And it's really all— it's about connecting people and allowing them to teach and learn from each other. A lot of the power is, is within that simple connection. It's not so much about the content. Content is important, right? But it's, it's a building. It's kind of like a that allows you to progress. You then need to actually put it into practice. You need to talk to other learners learning that. You need to talk with experts and teachers along the way.

Yeah, I think that's, that's so cool because thinking about the way that humans as a species have developed technological advances, whether it's, you know, oral communication or literacy or, you know, these abilities to connect therefore share and come together in this kind of collective intelligence. So harness that collective intelligence by externalizing the, the internal stuff that we are thinking and feeling and be able to share them and therefore also teach them. And I think thinking of technology in that sense as a mediator of relation and a mediator of information that can be shared and, and bringing people together, I think that's really interesting because that's That's always the way that technology has been used in the sense of technology most broadly construed, right? Yeah.

And I think we want to help institutions move in that direction. We want to discourage them to think about these technologies as simply a method to maybe boost their current systems and their current processes and optimize their current ways of doing things. We'd prefer to help educational institutions and individual educators and learners directly. Out to rethink what education looks like and feels like and give them new opportunities.

Yeah.

And, you know, in a kind of broad sense, you hear people saying, well, it's just a tool. We can use it as just a tool. And, you know, that's— it's a very, very powerful tool. But as you're kind of intimating, that is what we need to do with it. But we can also use it as a tool to help us transform the systems that we're in as well as just a tool to augment the current practice. You know, I was talking to teachers who were using it, you know, there are very clear ways that they can just augment their current practice, but there are also ways we can use it as a tool to actually potentially radically transform. Yeah, very cool.

One of the, one of the interesting stats that, that keeps coming back is 82% of content in boilerplate courses, the type of courses you get on most edtech platforms, even at the biggest and best schools in the world, 82% of the content you learn is irrelevant to the learner's current needs, which means that by the time they do need to use what they've learned, they've probably forgotten most of it. And we can all attest to this. Like, how much do you remember of what you learned at college? Or even, even if you just finished college, how much do you remember of what you learned last year? Probably not a lot, right? And that represents a huge inefficiency in the system. And it's because the system is built on the premise that we need to spend several months and years to of courses and then we are gonna shove kids in this course and they're gonna learn it for the next 4 years, then they're ready for, for the world. That way of thinking of education is fundamentally not suitable for the world we live in today. So it's those challenges that we are able to now try and solve.

Yeah.

But I think one of the interesting things that you are doing as a kind of external observer is that there is still a place for intentional somewhat abstracted learning of content.

Absolutely.

Right. And this I think is where some go one way and it's just all about content and then some go the other way where it's not, you don't need content because content's everywhere. Content's on Google. You can just Google everything on your smartphone and students using ChatGPT as a platform, just, you know, write the essay for me. Great. There we go. And then I can edit and change. And, you know, there is this middle ground of actually something, it is really important important for us, all of us, particularly for young people to spend some time with somewhat structured and mediated content so they can learn important concepts. They understand which concept's more important than another. So there's a kind of directing of attention towards certain chunks, if you like, that I think is, I think most educators would agree is still really important. And that's not always part of the tech kind of conversation.

Yeah, the way I like to think of it is, is that at the moment, educational solutions are kind of like you, let's say you, you're in New York City, you're in Brooklyn somewhere, and you say, you know what, I want to get to Central Park. At the moment, the educational system offers you bus tour routes that don't really start where you're at and don't really take you to where you want to go, but it generally starts somewhere close to where you're at and maybe takes somewhere close to where you want to go, but you first have to get to where the bus starts and maybe it's a crappy bus and you maybe have to jump to a different one. You know, that's the way that education works at the moment. And the tools we're trying to provide to institutions is, well, don't just provide students bus tours, give them Google Maps, give them, sure, here's a map of all the space that lies between Brooklyn and Central Park where you want to get to. And sure, here are some recommended bus tours that we have for you. But you can hop on, hop off whenever you want. You can start learning from exactly where you are and you can— maybe you want to hop off and go to the coffee shop or whatever the case may be. But giving educators the ability to provide students a Google Maps type of experience instead of just offering them predefined bus tours is the way that I think we can change it for institutions.

Yeah, that's interesting. Do you have any concerns that it deprofessionalizes the role of the professional educator?

Short answer is no. The professional educator is fundamental to the learning experience. Like I mentioned, it's only 20% content. 80% of it is from the educator, from experts, from other learners, right? So their role is fundamentally important. However, the role of designing courses and curricula is something that I don't think we've done very well. And one quick story. So I did, I did the research. Research project. I worked with a bunch of music schools. These are famous music schools. They have famous musicians as part of their faculty. These are people who are top of the class. They play in front of hundreds of thousands of people and then they teach on the weekends and, and some days during the week. So I interviewed them. They thought I was interviewing them about their life story, but what I was actually doing was I was mapping out what they learned and how they learned and when, right from kindergarten onwards. Like what was it that they learned from the moment they started music until today? Right? So in the background, I was mapping out what they learned and when, in which order. Then several months later, I presented to them under a different pretense. I presented them, hey, listen, listen, here's another course. Here's a new course we're gonna sell at the, at the schools. What do you think of this course? The majority of them actually became angry, say, why are you wasting my time? Obviously, this is not how you learn. You have to learn this before you go there. And then this whole structure is like, it, it doesn't make any sense. Why are you wasting my time?

Time.

And I think that highlights something, is that we learn in a very unique way. That is, you know, we bounce around between ideas. Serendipity plays such an important role in learning. You know, we need to embrace these serendipitous events. And I think the analogy of providing a map allows for that because you can say, you know what, I'm on this pathway, I'm going to take the recommended pathway recommended by my educator. But today, suddenly I have this problem. I have a real-world problem. I need which means I need to learn something that on the pathway is 7 months down the line, but I need to know it today, right? Because it makes sense to me today.

Yeah.

So let's allow the student to learn that concept today when they need it. This all becomes possible.

Interesting. But I suppose the, the advantage of the map though is to also be able for them to be able to track back almost and see where the thing that they've jumped to connects or different ways that it connects. Because it doesn't obviously connect in one way. And that's one of the advantages of some kind of a 3D map is you've got all sorts of different routes that you could possibly take. It's not linear, but at the same time, there's a connectivity that it's important that young people or that we all understand in our learning that how does this piece that I've just learned connect back and fit in with my previous schemas that I've developed, my previous learning that I've done. Yeah. But just to come back to the educator point. Because one of the things I guess I'm seeing as a possibility with your tool is that it allows people who are non-professional educators to structure curriculum and learning sequences in a way that is based on all the things that you're feeding into your system. So, you know, it potentially democratizes the idea of curriculum planning and learning design.

Yeah, we would hope so. We would, be fascinated to look, you know, pretend we are 2 years down the road, we'd be fascinated to look at what are the pathways folks are taking to learn things. How different are they from the recommended pathways for how to learn those things? That information's gonna be really interesting to look at.

Interesting. So you then, that's your feedback loop coming back into the system design as you keep evolving.

Yeah.

And you mentioned that the map is actually more far more than 2 or 3 dimensions. So we work in a high dimension space. We're already thousands of thousands of dimensions deep. This is going to increase significantly. Whereas traditional education is built on a 2-dimensional taxonomy, which, which is, I mean, it's worked for a number of years, but it's, it's no longer really that viable. So imagine mapping the pathways in a multidimensional space, really getting into the intricacies how people learn and what connections are being made. That's gonna be fascinating. I'm so fascinated. I'm gonna be so interested in exploring all of that.

Absolutely.

Yeah.

Interesting. I don't, I mean, I don't even know what that means, quite honestly, to go beyond 3 dimensions, but like, I can feel that there's something really important there because of the kind of the jumps that we make all the time, like the abductive jumps.

Yes.

These kind of hunches and these intuitions and these You know, they're far from linear and 2 or even 3-dimensional. Yeah. Interesting.

It actually, it actually highlights the difference between how we learn before we go to school and after we go to school compared to how we learn during school, because before and after school we do bounce around and there's no clear pathway that we follow. And yet we learn a lot. And yet when we go to school, it's, it's a very different feel, a very different path that we take to learn. And I totally understand why we needed to do that, but now we no longer need to do that.

Hmm.

Interesting. And I mean, I touched on a concern around the implications for educators, but do you have any other concerns, you know, that this whole ChatGPT and OpenAI kind of, and the potential for AGI and general intelligence, do you have any concerns more generally, more broadly?

Yeah, the worry I have with AI tools is that we are not thinking broadly enough on the implications of these tools and the potential benefits that it can provide. We're looking at it through the lens of, you know, we have these current systems and let's try to optimize our current systems. Whereas with AI, we have the opportunity, unlike with previous technologies, where we can actually start to rethink things from the ground up anew. And we could overcome a lot of the challenges. It's kind of like thinking of the Leaning Tower of Pisa. It was built over 200 years and they, they understood that it was, that it started to lean after the first 20 years. And yet people just carried on building and optimizing on top of the Leaning Tower of Pisa, right? And we don't need to build on top of the, the Leaning Tower anymore. We can think of, well, let's put the Leaning Tower to one side, but let's think of, okay, you know what, we understand the challenges. We have 150 years of the current educational system. Let's rethink that instead of just building on top of the 150-year-old leaning structure that, which is gonna collapse at some point. Let's build something new.

And I mean, I'm, I'm hearing the same call in all sorts of interconnected sectors of society, right? In the economy and in, you know, in the way that we deal with energy and all sorts of ways that this could really be a radical inflection point for everybody in all sorts of ways. Yeah, it's a fascinating moment.

Yeah, it's scary and it's got huge, huge opportunities and benefits in front of us.

Yeah.

Are there any ethical challenges that you face with it?

What is truth? How do we define that? There's so many opinions on even what is true. The way we're approaching it now is we need teachers to be in the loop. They are the ones that make that that final decision. It can't be the AI. We need an educator to, to make that final judgment. We're going to support and help as much as we can, but that's the way we're approaching it now. I think this is going to be a huge challenge moving forward because we're going to need to— we're going to have systems that potentially could be more intelligent than us very soon. They would be able to provide far more guidance on what is actually true. But how do we know what is, what is accurate? We're going to need to understand how the models work, how they think, how they come up with these conclusions. They are going to need to provide us evidence and help us understand what is true and not. But until that point, I think we're going to rely on experts to provide us that guidance.

Yeah.

That's the approach we take.

Yeah. Because you've got hallucinations confirming hallucinations if you're using AI to confirm. Yeah, it is.

But that highlights just, you mentioned something that highlights quite a key differentiator on how we deal with things. So So we do use large language models, but the input we have is the documents that educators provide us. They provide us the ground truth that limits the ability for the AI that we use to hallucinate. So we use the ground truth that the educator provides as the foundation.

Interesting. Yeah, because there was— I was thinking as you were talking, there is a challenge around the kind of authoritative knowledge that you get from this seemingly confirmed structured curriculum that someone's giving you, you know, there's a sense of authoritative knowledge built into that in kind of implied ways that's like, well, someone's taken the time to carefully build out this curriculum. Therefore, you're not questioning how true it is, you know, most of the time, whereas with a much more adaptive and much quicker turnaround of of taking information and turning it into this kind of structured learning design experience, you then have a potential challenge around, it seems like it's structured and interesting, but actually it's based on some fundamental misinformation.

Yes, it's, it's going to be a crazy time over the next few years before we have a handle on, on this misinformation is going to be a huge thing.

I'm Tim Clement-Jones, Lord Clement-Jones. I've been in the House of Lords for 25 years now. But what I did back in 2016 was set up the All-Party Group on Artificial Intelligence because I thought that parliamentarians didn't know nearly enough about the implications of AI. And then I was asked to chair the Select Committee on AI. Which reported back in 2018, and we did a subsequent follow-up report in 2020. And that report, both reports were really designed to look at government policy on artificial intelligence, looking at the opportunities, obviously, but also the risks, the question of whether we needed to regulate, if so, how we needed to regulate, what the ethics should be. And then subsequently, I was asked by Sir Anthony Seldon Priya Lakani and Professor Rose Luckin to chair their advisory board for their Institute for AI Ethics in Education. Again, it was the question about the opportunities. I think we were all enthusiasts about the application of AI in education, but it comes with quite a but. Not enough people, when they procure AI applications and systems, really understand what the risks are, what they need to do to mitigate them, particularly in in the education sphere, whether in terms of data or transparency and so on. And we produced a report back in 2021 now, which set out a kind of framework for AI in education designed to help those in the education field really kind of get to grips with the parameters that they need to set around AI in those settings.

That's brilliant. Thank you. And there's something that quite substantively has just changed in relation to— I mean, the way that educators generally, I think, are aware of AI has significantly shifted in the last few months. So obviously you've been doing this work for the last decade at least, but something significant has shifted recently with ChatGPT and generative AI. And what— so what's your feeling about the way that that has changed the landscape?

Well, I think it's very easy, and I've been looking at this for quite some years now, 7 years really, So I'm a little bit suspicious of situations where people have a kind of moral panic because they suddenly think, oh my goodness me, you know, what is happening? This is a frightening form of AI or whatever it may be. Or, you know, there's a risk of fraud and so on and so forth. I think it's really, really important not to react, but to look very coolly again at the opportunities and risks. And I think that things like generative AI, ChatGPT, GPT-4, GPT-5 coming down the track, They are opportunities in many ways. You know, what we got to do is understand what they represent rather better. So for instance, it's very good at summarizing a set of data, if you like. What it's not very good at is accurately delivering an answer, rather like a very inaccurate form of Google. So you've got to actually— it's horses for courses, but it could be a very easy and lazy way of passing exams. So there has to be a great deal of transparency about the use of these tools, if you like, by students, by teachers. You know, it has to be a common currency. But I don't think we should start getting too worried about this. So now, in a sense, the playing field has become a bit more level because we were talking in the Institute for Ethical AI in Education We were talking there about procuring by the school, by the teachers who were going to then introduce AI into the classroom, into the back office and so on. Now here we have a situation where students actually have access. You know, you can just go on the OpenAI website, as you probably have, and you can find out all sorts of answers to problems, some of them accurate, some of them inaccurate.

Exactly.

But it's a very, very open and accessible tool. And of course, there's a professional version, which no doubt will become much more used when it's more accurate in business and so on. So there's a lot of opportunity here. But we've got to be absolutely clear about the context and how these tools are used.

Yeah. Can I just ask you again, just briefly about that moral panic idea? Because it is easy to get swept up in the hype and there is a lot of hype and it's just, you know, there's a lot of conversation about this in my networks in education. Do you feel that there is something substantively different about what's just happened recently, or is it just an increasing level of awareness that has come to teachers? Because as you say, the young people even themselves are kind of bringing this problem or this opportunity into the classroom with them.

I think it's a big dose of both because the technology is improving by leaps and bounds. I mean, these deep neural networks are increasingly sophisticated. And, you know, when you look at what ChatGPT and GPT-4 have been trained on, it is a massive amount of data. The kind of exa-computing that is out there now available to people like Microsoft and Meta and so on is enormous. Now, you know, this means that it's only for the big tech I mean, if you look at Google's Bard, you know, they are based on huge databases which only they have access to and only they have the computing power to deal with. So big advances are being made very, very quickly. And as a result of that, you know, some very neat tricks, basically people answering questions, people doing amazing articles which look as though they're written by a human being and so on. The public has woken up to this. And of course, you know, every day an article is published. Only yesterday in The Times, they talk about AIs talking to each other, AIs having a simulating a social existence, which of course means that the metaverse is now going to be very much an AI domain where humans will enter into a metaverse which will contain a lot of AIs. And you may not be able to spot the difference. So, you know, again, there are dangers there, but there are opportunities as well.

Yeah, interesting. And I mean, as chair of the Council of Queen Mary's University as well, what's— how has Queen Mary's responded to this set of kind of opportunities and challenges?

Well, of course, this is the big thing at the moment. I don't think that there is an AI yet capable of spotting whether an AI has written something. Yeah, but that's what we've got to develop. So in a sense, we've got to develop the tech to eliminate the risks which tech itself creates. So all we can do is really rely on things like transparency, about honesty, and make sure that our students really do understand that this is something that we should embrace with open arms, but they should freely admit that they're creatively using things like ChatGPT. I mean, I'm in the course of writing a book about AI regulation, and I said, okay, tell me the 6 of the, the real issues surrounding misinformation online. And ChatGPT gave me 6 extremely good examples of what they were and how they could be dealt with.

Yeah.

And I thought, well, that's a very good start, but what I'm not going to do is have ChatGPT write the whole of the chapter. But it's very stimulating to have sort of response.

Absolutely. Yeah. And I know Reid Hoffman's written an entire book with ChatGPT or, you know, AI as his copilot, as it were. But it strikes me that it's a bit generic and vanilla. I mean, it comes out with, obviously, unless it's hallucinating and it has these mistakes, when it's accurate, it's good, but it's kind of mediocre and it's a good starting point. But the problem as a student perhaps is that, yeah, then you might take that and obviously submit it as your own work. But I agree, I think it probably won't be long before we've got more sophisticated tools to support with that. But I think I wonder how many of your, you know, the professors and lecturers at the university are also using it creatively as a co-pilot or co-learner, as a tool within the learning process.

Well, it's interesting because departments vary in their embracing of tech. I mean, you know, Queen Mary, we've got, you know, some areas where they're not particularly techie, but even in the humanities, you know, we've got digital music, we've got all kinds of, we've got a digital environment research institute, which is probably as we speak trying to devise AI systems in terms of assessment and so on and so forth. But there are a lot of people that I've met, you know, in quite recently in the last few weeks who are really involved in the whole development of assessment tools, which are exactly trying to find the difference between an AI-produced product, if you like, and a human-produced product. But it's incredibly difficult. The intellectual property implications are very great too.

Yeah.

And so just then, I mean, the main thing I'd love to ask you about is the government regulations, because obviously that's your main area of focus. And as you say, you're writing a book about it. It and have had a huge amount of experience. And I was reading your article about the recent AI governance white paper from the UK government and how you were quite disappointed by the response that the UK government had put out with that about the way that it was responding to the questions around governing AI. Because, I mean, is there an inevitability, firstly, that governance will just always be running to catch up with the tech because the tech is moving so fast and the innovation is much kind of incentive around the innovation that in a way, is it inevitable that the policy and the regulation will always be slower and running to catch up?

No, I don't, because I think that, and I think it's rather like the ethics of teaching, or, you know, there are some immutable values, if you like, that I think we should be aiming for. And that's why we need an ethical framework. But it isn't good enough just to simply say, 'Look, we'll let each individual regulator issue guidance on this, and don't worry, even though we've got standards, they won't be compulsory. This guidance will merely be to help you in your innovation,' if you like. I mean, this idea that regulation is the enemy of innovation is, to my mind, completely misguided, and I don't think the EU takes that view. And one of the problems is going to be that if we have AI developers, and we have many, many talented AI developers, they're going to find it very difficult applying the AI systems that they're developing if they haven't done the necessary steps to de-risk them before they try and exploit them in Europe and even in the States. The States is a highly litigious culture and the National Institute for Standards in Technology in the States is developing standards and sure as eggs, there'll be litigation around whether or not an AI has impacted on an individual and not adhered to certain standards.

Yeah.

And as a result, people will be very prone to being litigated against. Now, if we don't have some idea in the UK that there are certain areas where we absolutely have to have risk and impact assessment, and there has to be some overarching regulation which ensures explainability, transparency, accountability, freedom from bias, not purely related to data, but it related to outcomes and so on, and mitigating the black box aspects and so on, I don't think that we're going to be very well served. And we are then going to be playing catch-up. But I don't think that it is absolutely beyond the wit to have a broad sense of regulation which actually does anticipate new technological developments because you can make the ethics broad enough, you can make the risk and impact assessment broad enough. If you make it purely sectoral, then I think you're going to miss a trick. You know, I mean, for instance, is Ofsted really going to be the regulator for AI in schools? I can't see it myself. I'd much prefer the ICO, the Information Commissioner, as the sort of leadership regulator in that space to do that.

So you'd have a specialist regulator with the jurisdiction across different sectors rather than having each sector kind of regulate themselves somehow with, with the AI, but not really quite knowing and doing, doing it well.

Absolutely. I mean, there's no reason why the individual regulators shouldn't incorporate, you know, the overarching regulatory requirements into their regulation. So for instance, Ofsted, if if there was a general guidance from the ICO, might say, well, there are particular areas that we're going to focus on. And after all, we produced our framework for how that should be done when a school is procuring an AI system. So there's nothing wrong with, you know, looking at a particular sector. But there are some unregulated sectors too. You know, retail, for instance, loves customer data. But they're not, apart from, you know, the ICO, apart from GDPR, they're not really regulated at all. So, you know, that means therefore, you just simply let AI rip in certain sectors, which doesn't seem to me to be very sensible.

Yeah, interesting. And I mean, I suppose the Ofsted jurisdiction would be more around the way that it's impacting teaching and learning in the classroom. And obviously that, that makes sense that that would be their focus.

Yes, absolutely. And it might not impact on the actual students themselves. Yeah, you know, it's all about the school, isn't it? It's all about the teaching. It's not about necessarily about learner behaviour, and it's not necessarily about the back office either, for instance, and the capture of data and the outcomes as a result of the capture of data. I mean, Ofsted, you know, doesn't strike me as being a particularly digital regulator in those circumstances. And let me tell you, I looked at the latest DfE guidance which came out last week on tech standards. Well, it's all about the pipes and the plumbing. You know, it's not about the ethics at all. It's all about, oh, you know, this is what your broadband should do. This is what the standard of devices should be. And I thought, my goodness me, you've missed a trick there too.

Interesting. Yeah, there's a— I guess there's a safety in staying with the technical aspects rather than straying into something which is more about the values of how we use it, right?

Absolutely. And so in a sense, what they're already beginning to do is contradict this great idea that every sector has to sort of regulate itself in terms of AI.

And perhaps the ethics is the thing that often or always potentially gets forgotten or left till the last because it's not the most pressing materially, but it perhaps is the most pressing, you know, existentially.

Yes, absolutely. And there's nowhere more important than in education, quite frankly, because you're dealing with young vulnerable people you're scooping up potentially a great deal of data, the potential for inequity is very high. And it's at a time when the brain of a child, if you like, is extremely vulnerable, and is very malleable. So I think you have to take particular care.

Absolutely.

Yeah, that's right. And definitely, you know, the thing is that what are we expecting? Are we having a a sort of convoy situation where, you know, we move at the pace of the slowest regulator. I mean, that is again ludicrous. We've been waiting. I think the first government paper on this was 2017 talking about industrial policy for AI. And then we were promised, you know, ethics and so on and so forth back in 2017. So it's 6 years now almost where we've been waiting for some kind of certainty. In the meantime, Singapore's done stuff, uh, the EU in particular, the US is moving forward, and we don't associate the US greatly with regulation, but all of them have moved faster than we have.

Interesting. And I was listening to a few different podcasts talking about AI, and they were comparing the speed of the change with, you know, dog years, right? 7 to 1. And they're AI years are like 250 to 1. I mean, it's moving at such a phenomenal pace, the innovation and the development that, you know, you just won't be able to afford to stay, have this convoy approach and just take your sweet time.

Right.

And the really important thing is getting ahead of the game because design, you know, ethical design upfront is so important. And, you know, almost everybody who writes about this, all the, you know, people like Stuart Russell and so on and so forth, Mike Wooldridge, Dame Wendy, I mean, you know, everybody I know who writes about this, it's got to be ethical by design. What you can't do is retrofit. So unless you actually say, okay guys, upfront, this is the requirement, build this in, all you get is a whole series of black boxes. And that is unregulatable, basically. And so this is why people are so concerned about legislators and regulators running side by side. And it's not really a question of, oh gosh, these new bits of technology have come along. It is pretty clear what we need to do before artificial general intelligence comes along. Because if we don't do it now, it'll be impossible when AGI comes along. And we don't know when that's going to happen, but it's going to happen probably quicker than we thought originally. I always thought it was going to be at least 10 years before AGI But we're going to be seeing something quite similar because what you would do and what is already happening is you build all these different AI systems into a single system. So if you added, say, for instance, GPT-4 to DALL·E, various other forms of AI, because you've got every form of recognition, you've got language, you've got image, you've got, you know, calculation, you've got prediction, you've got it all under one roof. It's going to look suspiciously like AGI, even if it isn't quite.

Even if it's not. Exactly. And that's— and that it almost doesn't matter, right? If it seems like it is, it's a fascinating problem. And as you say, it seems that many people have been shocked by how quickly the capacity has increased. And therefore, yeah, we're probably going to be again surprised how quickly AGI comes.

Absolutely. And of course, the one word that I haven't used so far is trust. And, you know, the real reason and the reason we gave when we did our House of Lords Select Committee report, the real reason I'm so keen on making sure we have the right framework going forward is public trust. Because if you start a situation where the public are deeply suspicious of this technology, where they think their jobs are going to be taken, and of course, that's been one of the things that where the commentary has been very fierce. On ChatGPT and GPT-4. You know, people think their jobs are going to be taken. They look at the more mundane, more routine AI used by people like Amazon and so on, and they think, well, it's coming for us now. The professional classes think, yeah, lawyers, accountants, authors, painters, musicians, you know, and it's really important that we allay public fears by making sure that the liability and accountability and the ethics are absolutely clear.

Absolutely. I did want to ask you about that, just if I could, just finally. Are you hopeful about the kind of affordances that it will bring around things like jobs? Because obviously, I mean, clearly it's going to bring capacity that will and can automate huge amounts of, as you say, kind of more procedural work. How do you feel about where that will go and what the challenge for the labor markets, but also just the kind of general societal affordances of maybe just less work work, for example?

Yes, I think there's a great potential for greater productivity. That is absolutely true. But I think if we're not careful, we tend to focus at the moment on— we have the wrong emphasis for some of our education. I think creativity, critical thinking, all those kinds of human aspects are really, really important. Of course it's important. And, you know, we're not nearly numerous enough, more mathematical as a population. But that isn't the only thing. It isn't just an understanding of science and technology. I think what we have to do is understand, you know, creative potential as well. And therefore, the understanding needs to be about how we use technology in a creative way and embrace it in that sense. But you can't embrace something that you're frightened of and that you don't trust. So we have to build a framework that builds that trust. And that means we have to have regulation and risk assessment and so on. And at the moment, however cautious they're being, I'd prefer to be a citizen of the EU in those circumstances rather than a citizen of the UK where it all looks a bit live to me.

Interesting. Yeah, well, tempted to make a comment here about Brexit, but I won't. I'll refrain. No, brilliant. Thank you, Tim Logan. This is a really, really important and And it's, I mean, it's just a dimension that I wanted to bring into the series because the regulation part is often not seen as the most kind of interesting part. But as you say, I love that idea of ethical by design. I think that's really powerful and really important. So thank you for sharing. I really appreciate it.

It will be developed as this book takes shape. When it actually comes out is another matter because it's like hitting a moving target.

Yeah, I can imagine it'll be out of date by the time you've written the second chapter.

Absolutely.

Thank you, Tim Logan.

Great stuff.

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