Should We Still Teach Coding in the AI Era? An Interview with Prof Sue Sentance, Director of the Raspberry Pi Computing Education Research Centre at the University of Cambridge

0:00I do believe that in order to be a good user of AI, you need to understand something about the technology and how it works. That might be at different levels and might be just tiny, but that knowledge of AI at some level will help you be a more effective and critically engaged user.

0:27Welcome to Data & AI Mastery, the podcast where we bring you cutting-edge insights, practical advice, and inspiring stories from the leaders shaping the future of data and AI across the globe. I'm your host, Raoul Gabriel-Urmer, founder of Cambridge Spark, the leader in transformational data and AI upskilling, career development, and progression. In each episode, I will be diving into real-world case studies of companies harnessing the power of AI to drive innovation, reduce costs, and create new business opportunities. So whether you are an aspiring data scientist, AI engineer, or seasoned executive, this show is designed to give you the tools and knowledge to stay ahead in a world where data is transforming every aspect of business.

1:11Stay ahead, stay inspired, stay masterful. Welcome to Data and AI Mastery. hey Sue good to see you welcome to the show how are you doing I'm doing fine thank you and thanks for inviting me along it's a huge pleasure Sue as you know I'm a big fan of your research work I think computing education is extremely important I'm very passionate about it especially for the next generation so I can't wait to talk about these topics with yourself

1:48so to kick ourselves to you i'd love to you know for you to tell us a bit about your your journey right like why computing education i started off doing a phd in ai education would you believe it in which i got in 1993 so we're talking a long time ago and as part of that i developed intelligent tutoring system back in the day for language learning but then you know did my postdoc and went into school teaching and I taught in school for 12 years, schools and colleges. And my whole perception of education compared to what I was thinking about as a sort of developer, sort of researcher changed. I realised that what I'd done for my PhD was probably completely ill-informed and how important teachers were in the, my whole perspectives on learning and teaching really changed.

2:39And then I was around at the sort of the right time in a way. I, when the computing came into the curriculum in England in 2014. I was by then back in academia working in teacher education. And I had by then a master's in education, a PhD, you know, PhD, et cetera, in sort of AI, computer science. And so I got involved in research in computing education at a time when nobody really was doing this in the UK. And, you know, I've been very fortunate in working education departments, working at the Raspberry Pi Foundation. I'm now working in Cambridge in the computer science department and I have a small very focused RECUS research center and have a wonderful team of people looking at the many things that we can research in this area.

3:28Fantastic and as you know I'm a big fan of your research group having recently donated a scholarship so can you tell us a little bit what's the the current research interest? So we're now yeah in our group we're looking at we are looking a little bit of AI education and that's around you know sort of concepts and skills that you would need to teach about AI we've also got a sort of an area of programming education so how you teach programming I've got a PhD student looking at debugging and also some of the team that are sort of linked to mine and Raspberry Pi Foundation looking about how you might use generative AI in learning programming and then we have a we work with the microbit foundation and we're doing a longitudinal project on physical computing and then various other bits and pieces.

4:18That's amazing.

4:24Maybe to kick yourself more generally, you know, it feels like we live in a world that is moving really fast, you know, like technology adoption is happening faster than ever. ChatGPT came out a couple of years ago and it's already got over 800 million active users. you know, from a consumer point of view. So that's pretty intense. So in this age, I'd love to get your perspective of what are the skills you feel are important, you know, today, if we can call this the age of AI. So I think there are lots of skills that young people will need. And it's actually quite difficult for, you know, schools and teachers to sort of unpick those.

5:06But if we describe them in sort of buckets, there's, you know, young people need to sort of an awareness of AI systems. They need to know where it's being used, whether they're interfacing with an AI system, whether that, you know, what's happening to their data. Is it being used? You know, are they happy about it being used? Where is AI not useful? All those kind of awareness skills there's another bucket about being able to use ai so being able to to you know confidently use ai for the things that you want to do not just because the technology is there but as a an assistant you know for you so that involves being able to understand what you do want to do so you know problem solving and and that sort of thing and then you know the ability to sort of actually adapt quickly because it changes all the time so learning about it yesterday is no good for today so being a sort of you know a good user of ai and then my third bucket would be i do believe that in order to be a good user of ai you need to understand something about the technology and how it works that might be at different levels and might be just tiny but that what that knowledge of ai at some level will help you be a more effective and critically engaged user interesting so some fundamental knowledge of how it works the systems be able to engage manage the outputs maybe design with ai those are kind of broadly skills you feel are important you also said learning agility so how do we how do we learn how to learn faster I guess that's maybe an interesting question I really sort of pull back from the sort of the faster learning thing that's that's almost a separate topic in that we don't learn fast as humans do you know I mean we need time to sort of think and reflect and you know form our own opinions and that's the purpose of education and in a way this push to use AI to be more productive to cram more in can be not effective because we just end up with lots of stuff that we haven't really thought through so that's that's a sort of a side bit of your question the thing about actually that agility thing I think it comes down and it's for teachers as well as young people you know that that confidence that it's just a system do you mean it's just it's just another tool it's a piece of technology doesn't need to lead you do you know i mean it it's their business to help you and but we call that self-efficacy and if you lack that confidence then the next you know you you will be sort of thrown every time you know the system changes you know something's different and everything's moving so fast it's almost like the most important thing is that we have this self-efficacy and agency around new technology that's great that's really interesting especially with children right when you're developing it's about learning the fundamentals, be able to kind of like critique and adapt and take your time to form your own opinions.

8:32I really agree with that. Now in the employment world, things move really fast and there's a bit of expectation that you have to pick up the latest really fast in order to add value back to your business or as an employee. So that's an interesting dilemma perhaps to work out. Now if I take you to another sort of question is, why should we even teach AI and how should we teach it? Yeah, well, that's a big question. Really big one. Just we did a little bit of research a couple of years ago when we were surveying young people between the ages of 11 to 16, most of them between 12 and 14, about their confidence with AI.

9:18And really, we asked this question, what do you think AI is? Most of the respondents were just under 500 you know clean responses uh were from 12 to 14 year olds and it was just really striking how young people didn't really have a language to talk about you know AI they actually didn't know they sort of have got something in their head about it's you know it's there it's everything it's I'm interfacing with it but not able to say anything about sort of you know what it what what we meant by AI, sort of a lot of, oh, it's a bit like the internet. And then these kind of also, these more worrying misconceptions around, it's sort of, it's got human-like, it's like a human, you know, it thinks.

10:08I, trying to think, I had quite a good, AI is a non-human sort of sentient with a mind of its own, you know. And when we see AI presented as very human-like. We call that anthropomorphisation. It's leading young people to think that AI is like a human and also that it's sort of magical and that we sort of, you know, bow down to the technology of AI. And the reason for teaching, you know, carefully about what AI is and what it can do and a little bit about maybe how it works is so that we have informed, you know, that when young people leave school, they can you know they they will make good decisions even if they don't go into to ai at all because it's going to interface with every area of their life and if you sort of fast forward 20 years think of that child that's 10 now you know what is it that we we need them to know at school so is it about being in control and not depending on systems that appear magical yes i think there's there's like I said before there's the confidence side and then there's the ability to be to think critically about you know where and why AI is being used this is coming up in I mean there's quite a lot of new frameworks and competency lists and that sort of thing for AI now I mean it's obviously something that's of great interest to education systems and policy makers all around the world but something that is coming through at the forefront is this kind of idea of having being able to critically evaluate benefits and risks of AI yeah that's great speaking of frameworks so the the world economic forum has like a draft framework to define AI literacy and to think about in four sort of vertical engaging with AI creating with AI managing AI, the output and be able to challenge it and then designing with AI and, you know, problem solving and creating like some sort of system utilizing AI technology.

12:20How do you think about AI literacy? You know, everyone is creating a framework right now. The EU introduced AI literacy as part of the AI Act. I'd love to get your perspective of what do you think is important when we think about AI literacy for children? and do you think it's the same for the general population maybe for adults as well so yes we are I can see a good evolving of these frameworks so UNESCO produced a long list of competencies for students I think only 18 months ago maybe even shorter and also for teachers and you know It's quite overwhelming, long, large grid of things. And then this OECD framework, which I think is the one that you're referring to, I see a definite evolving in these sort of frameworks as, you know, every sort of, you know, new iteration sort of comes out.

13:19So the OECD definition of AI literacy representing technical knowledge, durable skills, future-ready attitudes, I think encompasses quite a lot. What I feel like these frameworks sort of maybe miss is that, you know, AI is used in context. It's used in a certain domain. You know what I mean? You can't critically evaluate the pros and cons of AI just like that. You can't say, now children, we're going to have a lesson on. And it's all dependent on where it is, you know, what you're doing, what the context is. So it becomes very interdisciplinary and very much attached to the context that you're talking about, the application you're using, or the purpose for using it.

14:10So in a way, the danger of some of these broad, encompassing frameworks is that they don't answer the question of where are you going to teach this? The statement that in this framework around AI literacy is like four lines long, but absolutely huge. And it raises a lot of questions. Really interesting. so if you're a parent today things are a bit overwhelming because on one side you want to support your you know your kids to to grow and develop and you can clearly see ai being used everywhere but you can also see that ai be used in in bad ways and you can't really trust it on the other side you can also see your kids using ai probably more than you are on a day-to-day because, you know, it's part of, you know, your childhood now.

15:00So what sort of advice would you give to parents in this context? That's a really good question. And we have been doing a little bit of research. We haven't finished analyzing yet when we've been interviewing parents and their 10-year-old children about technology in general. And we do find that exactly what you just said comes up, that parents might be some parents not all parents might feel quite anxious about what's safe for their young person to learn and how much should they be using to to use sorry how much should they be learning you know using AI and you know they don't really understand what they're doing on their screens and and and that sort of thing and wanting to protect their child and at the same time some parents have said to us that you know they they look to schools to teach young people this because they won't possibly be able to keep up and that their child will be able to teach them at the end of the day so obviously that's but even without having the the concrete knowledge I think it's the discussion with your child that is that is so important and going back to how we should be thinking more critically.

16:24In an ideal school curriculum, we would be spending a lot of time in discussion and developing those kind of thinking and arguing and questioning skills around technology. And unfortunately, currently, I don't think we do. We have curricular rammed full. So what can happen in the home, do you know what I mean, to complement what's happening in school is a discussion. You know, where is, you know, what's happened when you use CHAP-GTP, PT, what, you know, what does that mean? You know, so any discussion, a parent doesn't need to have the answers. It's just asking. I mean, critical thinking is essentially asking good questions and, you know, encouraging children just to ask questions and to be curious and, you know, think about where technology is being used.

17:18So it's great to hear your perspective, Sue. And I'd love to also get your opinion on what is sometimes confused. You know, there's this idea of teaching about AI and this is the idea of teaching with AI. So kind of encouraging teachers to use AI to create more personalized learning experience or for themselves to be more productive, you know, for example, for marketing and so on. Could you, from your perspective, tell us how you see the difference and what's important? yes good question so yeah so teachers can use ai for productivity you know so there's been lots of guidance given to teachers and there are lots of tools out there for teachers to help them prepare lesson plans and you know and they can you know speed up emails to parents and those sort of things on the sort of productivity side that don't directly interface with with with the young person so i think as teachers have a really heavy workload sort of training for teachers on how to do that effectively you know is is really good but then when we get to like teach teaching with ai for teaching and learning as opposed to teaching about ai which i think is probably the distinction you're you're getting to so yes so you're you can teach you know you can use llms whatever in your classroom you could help young people to do their research with llms and then for example my subject is programming so if you're teaching programming obviously the you have questions of when it's appropriate to use an am because obviously it can write all your code but it's producing an output that you maybe haven't learned anything from.

19:05So teachers need to make decisions about when with AI is relevant and helps the learning process and when it's, you know, when there's time for sort of thinking about sort of fundamental skills. And that's quite hard to do. So I think we're in a sort of learning journey there all around. I also think, so teaching about AI. so I'll just tell you about the way we divide up teaching about AI in our research center which is through the CME framework which is S-E-A-M-E social and ethical which is the one level which you can imagine that's all the sort of critical thinking stuff I've been talking about applications how you use AI models which is another layer down in a way of understanding some of the you know how how it works and then the engine level is really the the technical maybe coding you know that really sort of technical level and although we would develop this framework to be able to understand to have a language to talk to people about are this are you what are you doing at what point so if we bear that in mind we probably need to teach some of those different levels for depending on the age of students whether they're doing qualifications whether they're doing and and just focus on maybe focus on more of the social ethical and applications layers for students you mentioned programming and coding so you know i've been coding since i was six years old you know like my my mom and dad bought a computer i picked up html javascript and so on and then end up doing computer science specialized in soft engineering and obviously I'm biased you know it's helped me a lot it's helped me structured and problem solve and you know kind of like think a bit more logically I guess and rationally now in this environment today we talk a lot about how coding is being automated away right LLMs are super good at spitting out a lot of code and even you can create applications with thousands of lines in like seconds, which is, you know, quite a feat.

21:22So in this context, how important is it to teach programming? I mean, that is a much contested question. I'm not going to come down on one side or another. But the Razee Pai Foundation in the last couple of months produced a why we should learn, we must still learn to code positionality statement about how important it is still to learn to program and other sort of other people have really sort of spoken out in favor of code whereas you always hear learning to program whereas you hear a lot of we don't really any more junior programmers you know we can do all this stuff so basically we can use generative ai to you know we can use copilots generate code or whatever tool if their output is needed but we're not able to evaluate whether it's effective or not for the arguments for still learning to program are that you know it's we don't want to lose control of whether we understand whether that we can modify that whether can we modify that code can we evaluate that code you know can is it is it fit for purpose so to do that you need to be an experienced programmer you know we've got there's research that says that you know if you just get first year undergraduates just to generate their code for their assignments they're not able to tell you anything about programming or what it will actually do and so the focus on those undergraduate courses has become more on sort of code comprehension let's see what the code is and how do we understand how do we understand it and make small changes in terms of using tools the sort of current thinking like from the us is around teach the principles of programming just teach the basics so that you can and then use the tools to be more productive once you need to to generate code but then the the other side is well why do we even need programmers anyway so and then you get to these arguments of actually through programming you learn you know important skills like a decomposition being able to extract abstract problem solving having to have you know unambiguous you know solutions being you know these are these are actually you know you learn about logic you know these are actually important skills in their own right so i think there's there's various arguments to and fro you know on but on both sides but it's definitely I'm just going to say it's a debated issue.

24:01Absolutely. Thank you for your perspective.

24:09I'll take you to a quickfire round of questions if you're up for it. Yes, sure. Brilliant. Well, let's start with a contrarian view that you have. And obviously, we've talked about a couple actually already. but if you could maybe share one, what would that be? I believe that we should teach in a more interdisciplinary way. And we have very sort of, once we get to secondary education, we divide things up quite definitely into buckets. You have maths for an hour, then history for an hour, then computing for an hour. And then you might not have that subject for the next week. We don't teach in an interdisciplinary way and we should.

24:52And we focus too much on high stakes assessment, which of course is stress and anxiety and favours some particular groups of students. And those kind of ways of working, we don't do projects very much, are not helping us to young people to think for themselves. And I'd refer the listener to Gert Bieste's Three Purposes of Education to think about we're not just in school, just to learn facts. I mean, we are, you know, young people are developing as human beings. So we should be reflecting that in education. That wasn't a very quick fire answer. It's a great, great perspective. I really agree. My next question is, we've talked about Magic Wand before, but let's say you could solve any issue in computing education with a magic wand, what would that be?

25:56I think if we stick on AI, what we've been talking about at the moment, I think we urgently need professional development for all teachers in all subject areas, primary and secondary on AI the implications of AI I mean because teachers are just so important in the in the education system there's a lot of focus on things for young people directly but we we're not going to get anywhere unless we have a you know an educated workforce and it that needs to come from the top so my magic wand would be to the government to provide you know, comprehensive, free AI, you know, professional development for teachers.

26:47I agree. My next question is a bit more personal. Very curious about your answers. I'll kick off. What is your favorite programming language? Well, my PhD was in Prologue. No way. So, even though I really programmed only in Python for the last 10, 15 years. You did your PhD in Edinburgh, right? I did, yes. Yeah. Prologue is very, very big in Edinburgh. Yes. Yes. Yes. So that's my kind of like that kind of what brings out a sort of a warm sort of feeling inside. I love your answer. Brilliant. And what was your favorite subject back in primary and secondary school? Maths and languages. And that's when I think I got into computer science because it was that, you know, maths and languages, I think, fit really well together.

27:38and final question is what is your favorite music genre what gets you going I don't know about genre because I listen to all sorts my favorite band and for the purposes of this podcast I've taken my poster down favorite band is Arcade Fire I'm I love Arcade Fire I go and see them as often as possible I don't even know what genre you'd call it but that's that's what I'll stick with for now it's been absolute pleasure to have you on the show I'm a big fan of yours and can't wait to see your contributions for the field. Lovely. Thank you very much.

28:21Really delightful conversation. We've talked a lot about AI literacy, AI in education and the future. So let me bring a few reflections together. The first one is the importance of empowering the next generation children from primary and secondary schools to take ownership over AI systems. Think about the outputs, manage AI systems and be in control. If things go wrong, but also so we don't depend on this magical system for everything we do in our life. Now we discussed this idea of AI literacy in this context. And I like to think about it across four verticals, engaging with AI, thinking about safety, thinking about input and output of systems.

29:11The second is thinking about creating with AI, managing with AI. And finally, how do you design systems that use AI and might be combining multiple AI agents together to solve a problem? So there's kind of four components that one may want to think about in terms of AI literacy. And we had finally a really interesting debate about coding. You know, should we teach coding, the importance of coding to this world? And, you know, my personal view on that is programming slash coding teaches you really durable skills that are really vital in this environment. So critical thinking, debugging, root cause analysis, logical thinking, thinking in terms of steps, breaking down a problem in order to solve it, decomposition, etc.

30:00So coding is not just about building applications. It's not just about being able to edit code if things go wrong. It's really giving you a broad set of skills that you can apply in whatever projects and whatever problem you'll face in your life. So really fascinating discussion. Thank you for tuning into this episode of Data and AI Mastery. If you found value in today's discussion, make sure to subscribe so you never miss an insight from the leaders driving the future of data and AI. And if you're a data and AI leader looking to upskill your workforce with the fundamental data and AI skills to transform your business, Cambridge Spark is here to guide you every step of the way.

30:47Be sure to reach out to us on LinkedIn or on our website, cambridgespark.com. Until then, be sure to keep pushing the boundaries of what's possible with data And remember, mastery comes with continued learning and action. Until next time, stay ahead, stay inspired and stay masterful.