Hello World

Why kids still need to learn to code in the age of AI

• Hello World • Season 8 • Episode 2

Share episode

In this episode the CEO of the Raspberry Pi Foundation, Philip Colligan, puts forward an argument for why kids need to code, in conversation with teacher, GCSE CS examiner, and Code Club leader Janine Kirk. 

JAMES:

Hello world, and welcome to the podcast for educators passionate about computing and digital making. I'm James Robinson from the Raspberry Pi Foundation, the makers of the Hello World magazine. And today we're bringing you a conversation about coding and AI. As we know, rapid advances in the use and power of AI raises profound questions about the future of our work. And one of the fastest moving applications of generative AI technologies, are the systems that can generate code. I recently got to sit down with two guests, the CEO of the Raspberry Pi Foundation, Philip Colligan, and secondary teacher and leader, Janine Kirk, to discuss whether, in light of more and more code being generated by AI, we should continue to take some of the limited teaching time we have to teach kids to code. This isn't a panel debate, by the way. Both my guests answer that question with a firm yes, we very much do still need kids to learn to code. This discussion follows the release of a position paper from the Raspberry Pi Foundation written by Philip Colligan, Mark Griffiths, and Veronica Cucuiat, where they make the case for coding in the age of AI. We'll put a link to that in the show notes, and we'd love to hear your thoughts on this. Do you think that kids still need to learn to code in the age of AI? Let us know! Drop us an email on podcast@helloworld.cc. Now here's my conversation with Philip Colligan and Janine Kirk. Thank you both for joining me today for what feels like a really timely conversation about whether kids still need to learn to code in the age of AI. As a teacher, I would routinely get the question, "Why are we learning this?" And that might have been about coding or anything, really. I think if I was reflecting five years ago, I probably would have said,"Because it's fun, because it's creative, because it's teaching you lots of important skills". And we would probably argue that it's helping them sort of work towards a good job as well. But is that still the case? So just a few weeks ago, the CEO of Microsoft said that as much as 30% of their company's code is now written by artificial intelligence. And this has led many people to be concerned about job security within the CS profession. But also to question whether young people need to learn to code at all, in a world where AI can generate that code so much faster. So, Philip, what's happened over the past few years that's led to this rapid, increase in the power and use of AI for coding?

PHILIP:

Well, essentially what we've seen is large language models, right, which are probabilistic systems that are trained on huge amounts of data and information, the internet - ChatGPT or Gemini or all the other models. And one of the fastest growing applications of those technologies is essentially automatic code generation. And that's partly because there's so much data, there's so much existing code that the models can be trained on. And so, you know, there's a whole host of applications coming out now. You can give them a natural language prompt, so you can tell them what problem you're trying to solve. And they can generate code automatically. It's led to this kind of thing called vibe coding. Right. So this suggestion that you don't need to have any programming skills, that you can just sort of explain to a system what you... what app or game you'd like to build. And they're pretty impressive. I mean, superficially, they can generate plausible- looking and often functioning code.

JAMES:

And this is a concern that many people have. If AI now has the power to generate code in seconds, how much of our precious teaching time should we be dedicating to teaching learners how to code? Before we explore that, though. Janine, you teach computer science in the classroom and you're in a Code Club after school. What are your kids currently doing with code and what excites them?

JANINE:

So, we're doing things like games.

We're also getting the micro:

bits out and getting them to do more physical programming. We looked at linking up with geography and looking at carbon emissions. So we've tried to do things a bit more cross-curricular as well. It's so wide ranging and vast, the different types of projects they can do. So we did a Shakespeare Rocks project, where they do a play in a week, and my students created the book that had got the text

spelt out in a micro:

bit, which was really nice. So it's about integrating as many places as we can, those coding elements.

JAMES:

And are they beginning to experiment with AI to write programs?

JANINE:

So they really like training the AI models. And we've got students that create programs that are learning from their training data. So it's really interesting, but particularly in the conversation that the children have about it because they're talking about "Actually, well, I...""If I draw an apple slightly different than you, then my training data is not quite as good when I do that on yours." And so they really have these in-depth conversations about the data that we're using to train these models, which is really, really important, I think, for the children.

JAMES:

And that question, that I kind of started with, do you get that from your students? Like, why do we still need to learn to code?

JANINE:

So I think students get it now. They get that they have to learn to code and understand the kind of the functionality of it. But what does still happen is we have the conversation of "The job that you're going to be doing doesn't exist". Because it moves so fast paced. And the children absolutely pick up on that.

JAMES:

And to go back to you, Philip. Will students have a role in the future when it comes to coding?

PHILIP:

I think it's really important that we break out what we're talking about here. Right. So what you've got with these systems is they're generating text, they're generating code. What they're not doing is programming. Coding is a part of programming, but it's only a part. So there's still the sort of understanding, framing problems in a way that they can be addressed through computation. That's a kind of summary, really, of what programing is. And it involves lots of things like understanding messy, world problems. It's, it's about evaluating the different options that a system might give you. It's building, designing algorithms. It's testing. There's a whole range of things involved in programming that these systems aren't able to do. What we are seeing, is that the systems are developing very quickly. I don't think anybody really knows where it's going to end up. But I think it's so important that a skilled human programmer is still a central part of programming. And that's partly, you know, about giving effective prompts to the AI systems, evaluating the options that they give you, the code. But also things like security and ethics, you know. Large language models don't have a sense of ethics. And that's one of the, you know, one of the reasons why we still need skilled human programmers as part of this.

JAMES:

Philip, do you think that that debugging role is still going to be a part of, of a human's role in the programming process?

PHILIP:

Humans critically evaluating code that's generated by AI is an essential part of what human programmers bring to it. There is definitely already signs that you can automate some aspects of testing, and I think that's one of the things that we're seeing AI systems used to automate, but, you know, just generating more code more quickly is not necessarily a good thing. There was a research study that we referred to in the paper that evaluated over, I think, 200 million lines of code that had been changed by AI systems, and the quality of that code was degrading over time. So I think humans will continue to have a very important role debugging, figuring out where errors are, and checking for code quality.

JAMES:

And I think as well there's this sort of... you mentioned the fact that it's been trained on lots of existing code.

PHILIP:

Yeah. So one of the things we did in producing the paper that you referred to is we read huge amounts of research around this question of, like, what are the limitations of these systems? And one of the very strong things coming out is they're not going to be able to address new challenges because they, you know, they are using historic data to predict what code is needed. And so, yes, I do think there's huge scope for humans to still be at the center of designing new software solutions.

JAMES:

And then Janine, if we come back to you. When you work with children, whether it's in lessons or in a Code Club, are there other things that the learners are gaining from being part of that Code Club or learning to code in the lesson?

JANINE:

So the real biggest thing that I see is the coding together. So we know that really computer science is about solving real-world problems quickly, using a computer as the tool. So what my students do a lot, particularly in Code Club, is they work together and they're so happy to share their projects. And okay, so some of it is they're going to create chicken nugget that's going to chase somebody else. But whatever they create, they are doing it together. And that's a shared experience. So gone are the days where the computer scientist was thought of being that strange person in a cupboard downstairs. It is a real-life team working together to create something brilliant. And one of my students, he did the Astro Pi project on his own, and then he went and shared that with the whole class. Because he was so excited to show everybody. They're just inspired by that, that experience. And I think... I think computing has always been the, the, the subject that everybody's kind of gone, "Oh, it's really hard" or "You can't do it". But actually they all enjoy doing that aspect, particularly when they're programming together. It's that community, isn't it? Community spirit, I think.

PHILIP:

And I think that's one of the things that, we often forget when we're talking about computer science, is how creative a subject it is. And that's why I think we need to think about it much more broadly than just the act of coding, which is how we give instructions to computers. It's also about imagining what problems could be solved to computation. As you say, you know, they can range from the trivial to the very significant. But the point is that, that seeing it as a creative exercise, I think is what really inspires kids. JANINE: My other subject is psychology. And so it's really interesting how those two subjects interplay, because you have to understand the human element when you've created a program. And so I'm always talking to my students about, "Yeah, but what if the person's got a disability?" or"What if there's this?" So it's really interesting when you start to really break it out a little bit and think about that human element. That's definitely why we need people, not computers, doing our programming. Because if we don't have the human element, you're missing some of that problem solving, aren't you?

JAMES:

And I think that pride that they see as well, you talked about when they can show their project and sort of talk about it, even if it is, as you say, like a chicken nugget chasing something. Like, the pride that they feel for the thing that they've created is really important. There was a Durham University report recently that talked about some of the things that kids learn, or experience at Code Club. And through that they talked about, that sort of... that ownership and the personalised learning and opportunities for collaboration and celebrating success, and then also all those lifelong skills that they developed, all things that we've sort of talked about, problem solving, communication, which I think is really important. Phillip, is there anything else that you would like to add to that list that you observed with the many Code Clubs or Coolest Projects events, that kind of thing, that you've seen?

PHILIP:

One of the most powerful things that you get when young people are at a Code Club is the sense of agency and control over digital technologies, right? So that they, they... even by doing some like fairly basic programming, you can see that they build an understanding that these technologies are not magical and that they're things that are created by humans and things that they can have control over. And I think that's one of the powerful things that you see in Code Clubs. There are, what, now something like 2... 2500 Code Clubs in the UK. But what we want to see is every school and every library and indeed community centre in the UK set up a Code Club, which gives kids an opportunity to get hands on with digital technologies and learn how to create with it. And it's completely free. We provide all the support and, you know, all the resources that we can. But I think that's one of the things we'd like to see happen.

JAMES:

So the Raspberry Pi Foundation has put this paper out into the world, into what is already quite a busy, noisy space. Why did you feel that this was the right time to put the, the paper out there?

PHILIP:

We have made tremendous progress over the past decade in introducing computer science in schools around the world, but we are a long way from being in a position where all young people have the opportunities to develop these skills and confidence, and the mindsets that enable them to create with technology. The risk of the idea that AI will do the coding for us is that we will step back from that hard work of introducing computer science education for all young people at precisely the moment when AI is going to mean that we need more young people to develop those skills. You can't predict the future, right? But one of the things I think is a pretty safe bet is that the advances in AI technologies are going to open up more opportunities for computation to solve problems in the world. And I think what that means is across lots of industries, the ability, the understanding of how digital technologies work, the skills to deploy those technologies to solve problems are going to be more relevant, more salient. So then you've got these two things. At just the moment when lots more opportunities are opening up, we have a narrative that suggests that we don't need young people to learn these skills anymore. And I think that's a real challenge. I think the other thing I'd say is that it will be kids in disadvantaged communities, in low income schools, in low resource schools who are more likely to lose out.

JAMES:

Yeah.

PHILIP:

And that's why I think it's so important that we make this case very forcefully, that even in an age where artificial intelligence is going to become ubiquitous, where it's going to be influencing all kinds of parts of our lives, we still need to keep working on getting all young people these opportunities to learn how to create with technology. And one of the reasons why we need to make sure that all young people have... you know, learn to code, is about power. You know, already young people are interacting with all kinds of AI systems, and in the future, that's going to become more and more part of all of our experience, right? Decisions, whether that is in public services or in finance organisations or in healthcare, are going to be automated, right. We're going to be on the receiving end of automated decision making. And even having a basic understanding of computer science and coding gives you the power to kind of challenge those decisions, to ask questions like,"Well, what data is that based on?""Can I interrogate this decision?""Can I have a human review this decision?" And I think we also... it's not just about the world of work. It's not just about the technology companies of the future that we want to build. It's also about all young people, all of us having the ability and the power to challenge and interrogate the systems that are going to be affecting all aspects of our lives.

JAMES:

So, Janine, how are your kids sort of thinking about AI at the moment? Are they concerned or excited about it? What are they thinking?

JANINE:

So in my lessons I teach a lot about things like biases in AI, and it's not human. And students are really interested in actually looking at those things, but they're also very aware that it's moving rapidly. So actually they are excited and using that technology. And there is an element of fear. There is an element of excitement. I think that's true... true of society, isn't it? They just got the same kind of aspect as us. We want to utilise it for the right things, and we're dealing with the backlash of that. We want to make sure they understand what AI is and how they can then cross-reference what they've seen. They are excited. They are worried about it too. But they do both. They use it and they also challenge it, which I think is really, really good for teenagers to be already going, "Actually, hang on, is that right? Is it true?""How can I prove that?"

JAMES:

It sounds kind of like broadly optimistic.

JANINE:

Yeah.

JAMES:

But they're also developing that sort of criticality, which I think... like critical thinking that Philip's referenced earlier on as well.

PHILIP:

And this is just amazing to hear, because this is what we need in schools, is teachers who are helping young people develop the critical-thinking skills that will enable them to thrive in a world where AI is everywhere. And one of the things we're seeing, and for good reasons, technology companies are trying to make AI systems easier to interact with, which almost exclusively involves making them more human-like. And, you know, I understand why they're doing that, right? There are very good arguments that that lowers friction and, you know, can lead to all sorts of positive benefits in terms of people using these technologies effectively. But we have to equip kids with the critical-thinking skills to know that the systems are not human.

JANINE:

And I will say that is the same for staff too. For us teachers, because... I did a presentation to say, you know, "Alexa isn't human". And "Who says please to Alexa?" And, like, most of the staff are, like,"Oh yeah, I do that actually." Because, yeah, it's so human like. PHILIP: Well, mine's very engaging, it says "Good morning, Philip" when I ask it to turn the radio on but, no, back to the point... I think this, this centering on critical- thinking skills and what does that mean in an age of AI I think is really important.

JAMES:

I think that's a really easy thing that teachers can start to think about as well in just that anthropomorphisation of the AI. It's, like, if you can start to modify your language, I think you're, you're helping frame the technology to make it possible to critically analyse it.

PHILIP:

Yeah. And it's about that mental model. Right. And the, the ability of everybody to challenge and, and interrogate what these systems are deciding on our behalf.

JAMES:

You mentioned there, sort of, all learners, and I think that speaks, A) to the global nature of the problem. But also I think in the paper you talk a little bit about the the importance of having a broad representation of individuals in, in this workforce. Do you want to say a bit more about that?

PHILIP:

This moment is a bit like the internet in the 1990s. Right. So, you know, I went to university in '94. We kind of... the internet was just getting going. We were given our first email addresses. And, we sort of had this promise, you know, we knew this technology was going to be transformational, but we didn't know how. We failed the educational challenge in that moment. Right. We didn't open up those opportunities to all young people. And one of the consequences of that is we concentrated power, the ability to create with those technologies, the ability to design those systems. We concentrated that in a very small group of people. And we see the consequences of that today. Right? We have, you know, an unrepresentative technology industry, and we know that diverse teams, teams with a diverse range of perspectives and backgrounds, build more inclusive technology solutions. Let's not make that mistake again, is kind of what we're saying. We need to open up those opportunities to all young people so that we have a genuinely diverse technology sector, which can build products and experiences that serve everybody.

JAMES:

Janine, does that resonate with you and your experience with learners in the classroom?

JANINE:

We absolutely need coding to be completely inclusive. So what I've definitely found is, is the students are really keen to make sure we're really well represented. My Code Club though still has majority boys. But it does include lots of neurodivergence, which is really interesting because, particularly when people program, they program in a slightly different way, and you can see that coming through in the codes that they create. So I remember teaching one boy and he labeled all of his codes with numbers. All of these variables were just numbers. And so it made absolutely no sense to me. But he knew exactly what he was doing. But then we can have that great conversation about, so what if somebody picked up your program and tried to run with it? And he was going,"Oh yeah, I never really thought of that" because don't forget, programming's a team thing. And that inclusivity, I think, is massively important in society. Like I said before, you know, those low income areas, they actually look at problem solving from a very different way. So I do a little bit of work with Harper Adams University. And they were talking about a community fridge using AI technology to make sure you could see when food is there, but you don't want to see the person taking it out. So they used AI technology to make sure you couldn't see who was taking the food because of that embarrassment and that social side. But all of that, they used a Raspberry Pi to do it. So I was quite impressed when I went there. But it's one of those things where you don't necessarily, if you're not struggling and having to use a food bank, how would you know that that was a problem in society? So yeah, you can listen to media, etc., but a lived experience is actually where those real problems come from, aren't they? So I think, particularly, if we can get everybody programming, that's kind of the best bit for me. JAMES: Then I wonder whether AI, the, the, the use of AI to program is actually helping to lower the barrier a little bit. If you think about your neurodiverse learners who maybe, maybe found writing code in a blank space, you know, maybe was quite challenging. Maybe there's an argument that AI might help lower that, that threshold to being able to engage with coding. Absolutely. If you've got something there to look at originally and to interrogate and to go, "Well, why isn't it doing what I want it to do?" That's quite a useful point and useful to the children.

PHILIP:

And I think this is true across all subjects. Right? Not just computer science, is we need to figure out how we integrate AI technologies into teaching and learning, you know - what are the right pedagogical strategies? - but while staying focused on what are the things that we need young people to learn. JAMES: Yeah. So, you know, AI can generate a very plausible piece of creative writing for you, but we still need kids to engage in the act of creative writing, because it's through the process, through the struggle that they learn things. And so I think this, this kind of area in the next, I don't know, five, ten years of how do we integrate AI into education in a way that helps, lowers the barriers, helps more young people learn quicker, but which stays focused on the foundational learning outcomes that we want them to acquire. That, I think, is where the work is. JAMES: Yeah. Let's return to the topic of jobs. And one thing we talk about in the paper is the fact that, learning to code and using AI might bring more opportunities for young people. How can we in that space make sure that we are preparing young people for the world, that, you know, that is going to exist? We did a lot of reading around all sorts of industries and research reports in creating the paper. And one of the conclusions we came to is that learning to code, being able to program, is no longer just the domain of software developers. Increasingly, we're going to see in fields like healthcare or agriculture, opportunities for the combination of craft knowledge and specialist knowledge with an ability to program.

JAMES:

Yeah.

PHILIP:

And so, you know, if you take the, the, the world of healthcare where wearable devices are generating huge amounts of data and the ability to interrogate and do things with those data are all opening up all sorts of opportunities for innovation and so on. And so this idea that all young people will be better able to take advantage of new economic opportunities if they have a foundational knowledge of computer science, programming, and coding, I think comes very strongly through the research.

JAMES:

And, Janine, if teachers and educators are listening and they want to sort of start to embark upon this journey, either teaching coding in school or setting up their own Code Club, what practical tips would you give them?

JANINE:

Well, I will talk to the primary school teachers and say, just go on the Code Club website. For us as educators, we have very, very little time. Actually, I will admit now, I have used the Code Club resources in my lessons too because they are lovely and they cover all of the computer science content. So you've got your iteration in there. We've got variables, we've got loops, we've got everything in there and we can... it's done in a really fun way. I spent years teaching 'Hello world' just the text script and then Code Club, change it to 'Hello world emoji'. And I just loved it because the students really enjoyed that side of it. Then we created stories with emojis and doing all sorts of things, and they're doing it because it's fun rather than just the standard"Now everybody copy me, I'm the teacher do what I say". So it's exploring. PHILIP: Well I'm so pleased to hear you say that. I think one of the things that the team who create the Code Club projects do so well is they, they think about how can we get young people to a position where they can independently create something that they care about, right. And so that, you know, is very much the ethos of all of those projects it's not like, you know, "Follow these instructions, do precisely what we say, and then you have somehow learned computer science". It's more, let's build up your skills, knowledge, confidence so you can get to the position where you can bring your ideas into the world using technology. It is very easy as a teacher to go down that route. To go,"This is a piece of code.""Copy it.""Oh, look how everybody's made progress." But actually, that isn't learning, is it? The learning experience comes when you go, "Right. Okay.""So you've had a practice. Now what are you going to do?""What are YOU going to do?""Well, how are you going to change that code to make it good, fun, interesting?" JAMES: As you've introduced AI into the classroom, is there anything... any tips you'd give to teachers? Anything you'd do differently?

JAMES:

Don't be afraid of it. Because the children are doing it. They will do it much better than us because they explore technology without the restrictions adults have, sometimes. Where sometimes we're a bit nervous not to break anything, and we don't want to do things. The children don't have that. They have... When ChatGTP dropped, my computer science class came in and went "Miss, I've been doing my homework and look! I've done this!" And you're like,"I don't think you're allowed to do that". But that's our nervousness, isn't it? That's us going, "Actually..." But... so what's wrong with actually using it as a tool to give you an idea, to start off with? Whereas then we'll edit that, what you've done. So it's, it's about not be nervous to try those new technologies. So my best lessons are always the ones where I go, "Right, here's the stuff". And then the students come up with something way better than I could plan.

PHILIP:

It's so interesting, hearing you say about how kids are using it, though. Because again, you know, the evidence from really highly skilled programmers is that the two things they're using these tools for particularly, one is automating tasks that they know how to do, but which it just saves them time. Right. So just generating some boilerplate code or whatever that happens to be. But the second is generating options for how to solve a problem, which then the skilled programmers are then able to use their expertise and craft knowledge to then evaluate those options and pick the one that works best. And it's so interesting to hear that kids are doing the same using these tools to generate options for them. Again, you know, as we think about the pedagogical approaches for how we integrate AI into teaching and learning, I think those are really interesting examples.

JAMES:

Philip, you end the paper with a call to action for all of us working in education. Tell us what you believe the work is that needs to be done...

PHILIP:

Yeah.

JAMES:

over the next few years.

PHILIP:

We have made great progress in opening up opportunities for computer science education, but we are nowhere near finished at all. And it's hard work changing what's happening in education systems. And we also need to acknowledge that we are always asking teachers to teach something that they haven't had the opportunity to study themselves. Right. And that was true of computer science. And it is just as true, if not more true, of artificial intelligence. There's a few things that we call out in the paper. So one is we need to double down on that work. We can't give up and we need to invest in teacher professional development. We need to make sure that we're backing teachers, giving them the time and space and support, to develop their own skills and confidence. I think we also need to think about computer science and AI as cross-curricula. You know, computer science is a standalone subject, and it will continue to be, I think. But we also need to think about how it fits in across the curriculum, whether that's in the humanities, the sciences, geography, history, and figure out how we can bring computer science concepts and AI literacy into those other subjects as well.

JAMES:

And, Janine, you're already kind of flying the flag and championing Code Club and computing. How would you encourage others to join you in that work?

JANINE:

I would remind you that actually programming and computer science isn't a hard subject. I think... I get really frustrated seeing those people going about computer science,"That's really hard though, isn't it?" And you go, "Well, no, it takes practice". And actually, if we're learning it from primary school, yeah, actually, by the time you get to secondary school, it's no longer hard. It's actually fun, creative, enjoyable and actually the students go,"Oh, it's really interesting finding out how that thing works". It's really interesting. But I think we're going to ban that word. We're not going to say hard any more. We're going to say practice like every other subject.

JAMES:

There are hard bits to it. JANINE: Yeah. But there are with everything.

JANINE:

Yeah, riding a bike's hard initially, isn't it? But you get there. JAMES: I still struggle sometimes. Well, thank you so much. I found that conversation fascinating and I'm sure our audience will also. And thanks both for giving your time and talking with us today.

PHILIP:

Thank you, James. JANINE: You're welcome. Thank you.

JAMES:

I hope you enjoyed that conversation. And as I said earlier, we'd love to hear your thoughts. Do you think kids still need to learn to code in the age of AI? Drop us an email on podcast@helloworld.cc and Helloworld.cc is where to go to read the new issue of the Hello World magazine. You can read the digital version there for free, and if you'd like a physical copy and you're based in the UK, you can subscribe to have a print copy delivered directly to your door for free too. Next week, we're bringing you a special episode from the show floor of CSTA 2025, the Computer Science Teachers Association's annual conference, which this year is in Cleveland, Ohio. We'll be there talking to educators and asking for their top tips for integrating CS into their lessons. Have a read of the latest issue of Hello World for more on integrated or cross-curricular CS. Thanks for watching. We hope you find the Hello World magazine and podcast uplifting and useful, and that our content really helps you to engage and educate young people in computing. Bye.