Does "vibe coding" have a role in programming eduction?

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 podcast. And today we're asking whether the term 'vibe coding' is a useful one in the context of programming education. 

The conversation is hosted by Jane Waite, senior research scientist here at the Foundation, and she's talking with three guests: a university academic, a secondary school teacher, and a student. Who have each investigated the use of AI tools to support students learning to program. Here they are to introduce themselves.

DAN: Hey everybody, I'm Dan Zingaro, I'm a teaching professor at University of Toronto.

IRENE: I'm Irene Stone and I'm a computer science and mathematics teacher at second level in Ireland, and also a doctoral student at Dublin City University. 

JEDIDAH: My name is Jedidah Ajala and I am a first year computer science student studying at the University of Edinburgh.

JAMES: Thanks, Dan, Irene, and Jedidah for joining us for today's conversation and thanks to you for listening. Do let us know your thoughts. Drop us an email to podcast@HelloWorld.cc 

Without further ado, over to Jane.

JANE: Thank you for joining me today. I'm really excited to be able to talk to all of you. This is a wonderful opportunity for us to, to think about generative AI tools and maybe this idea of vibe coding and how they're changing the way that our students learn to code in classroom contexts. And I'm going to start by asking Dan. 

Dan, what is vibe coding?

DAN: Hi, Jane. Hi, everybody. So Andrej Karpathy introduced this term. It's, by now, I think I could call it a legendary tweet. It was in like February 2025. I think one thing I want to start with is just to say that the meaning depends on whether we're talking about a professional developer or a learner. 

It's a definition that's currently being negotiated, I think. It, to me, feels different than the other types of AI-assisted coding we've had before. It's not quite agentic coding because generally it looks like the human is in the loop for vibe coding. And they're not typically in the loop for agentic coding. 

But anyway, sorry, Jane, you asked what the definition is. So when, when professional developers are vibe coding, what they're doing is, they're creating prototypes or they're experimenting with prompting, and they're not looking at the code. Okay. So you don't review the code, you don't test it. When there is an error, you paste the error message into the AI and ask it to fix the error. If you get really stuck, because we're talking about professional developers right now. If you get really stuck, then you can look at the code. Because if you're a professional developer, you know what the code is doing. And this is where I think vibe coding is quite different for learners who don't have that expertise to fix it when something goes wrong.

JANE: I completely agree, and it is entirely different, isn't it? That we're talking about children learning to program with the help of AI. Whereas when we're talking about vibe coding, as Dan said, often it's associated with being a professional developer. And even though they might occasionally talk about using it with students, that's not something you tried, Irene.

IRENE: When it comes to vibe coding, no, I've never used or done vibe coding. You know, students really need to understand how code works, you know? So it's not about generating working programs. If they can't read or write simple code, then they're not going to be able to evaluate any AI outputs.

DAN: I agree, and I think there's so many posts right now online that make vibe coding look easy. Like it just, it looks like, oh, you know, just just do whatever you want and ride the vibes, you know? It's like, well, that's easy for the professionals to say, right? So we need, we need to be careful with that. Yeah.

JANE: Yeah, I agree. And actually, Jedidah, it's not very long ago since you were doing computer science in school, learning how to program. And I just wondered, you know, from your perspective as a student, how much did you and how much your peers using either vibe coding or other uses of AI in learning to program?

JEDIDAH: I would say in my case, in the case of my year at school, I just finished S6 and I'm at first year of uni. AI, especially vibe coding, is used minimally, especially in the exam project environment. So for example, in my Advanced Higher projects, project development was very linear. You go through your analysis stage, you get your requirements out, and then you design based off of those requirements, you implement, and then you test and evaluate, you know, your clear-cut waterfall method and it's cut in stone really. 

So at school, that was really the only process that we were learning or exposed to. Well, we did learn about agile methodologies, but in terms of traditional context, we learned the basic traditional ones and AI, depending on your teacher, was only used in these contexts. If maybe you wanted to brainstorm your requirements or check if a test case was correctly formatted. So it was used with instrumental value rather than as a starting point for a project that you're doing in school. 

Outside of school, though, it's a little bit different. I would say if you've got a project, project that's going on or alongside development that you're trying to build on your own, that's when AI models and something similar to vibe coding is used. 

For example, I went to my first university hackathon at the weekend, and I think that was my first interaction with vibe coding and kind of like what Dan says, online everyone's saying like vive coding is easy, is laid back. It's fine. But, I was in a team of four girls and it was our first time being full stack, no, 'postdoc developers', under time conditions. So that's where we employed Claude Code and other chatbots to try and get to our solutions. It was our original idea, but we took our knowledge, we told it we know Python. We know SQL. We don't know React. We wanted to make an app, but we made it into a website instead. And then our development process was guided using vibe coding. And as developers with very limited knowledge, we could still tweak the code that came out from the model and things like that to make a fully functioning solution at the end of the day. So I think that was the only reason it was possible within the 12 hours we were given.

JANE: But Jedidah, you've already done A level computer science. You can already read code well. And were you with other girls who were similarly computer science undergrads?

JEDIDAH: Yes, we were all first year having gone through the whole school system and in computer science as well.

JANE: So you, I would say, you were already junior early developers and that's really different, I think, to coming completely cold as a, you know, a much younger student who's not started. But what's really exciting is that your A level has got you into that point where you can have a go. I think that's just so it's just so wonderful. I'm so proud of you. It's so good! 

So anyway, let's kind of move on a little bit. And I just wondered, actually, let's just ditch that term 'vibe coding' for our chat today. And maybe we should explore what we might teach and how we teach that when we use AI in the teaching and learning of programming. And Dan, I know you've done a lot of work with Leo Porter, but you've kind of been looking at when to introduce AI into the kind of the process of teaching programming.

DAN: Yeah, we've gone on a bit of an adventure, in the past few years with your question here, Jane. So early on, Leo and I wrote a book, teaching programming with GenAI. Definitely don't look at the first edition of that book. We made some mistakes, and I'm sure we've made mistakes in the next edition, too. But, we don't know what those mistakes are yet, so it's TBD on that. 

But, in the first edition, one of the mistakes we made was introducing it immediately. So it was like we had this exuberance, Jane, you know, where there's something new? And you're just like, wow, I like, you know, this is going to be so cool, right? 

So chapter one was... In a single chapter, it was installing VS Code, installing GitHub Copilot, writing your first program, getting Python to work. And Leo taught the course based on the book. And he was like, Dan, we got to, we've got to slow it down. Like, students don't know what Python is doing, what VS Code is doing, what Copilot is doing. What's AI, what's not AI, so we dialled that back. 

It makes sense to delay introducing GenAI. At least to not do it immediately, so that students are aware of which components of the programming stack are doing what. That's what I see now, although I'm imagining in future, separating these components is going to become very difficult. For example, you know, much to our disapproval, students don't necessarily know the difference between Python and, for example, the IDE that they're using to run their Python code. And it's totally understandable that they would have this misconception because they're, they're so integrated. Right. And I'm imagining the same thing is going to be happening with AI, where we're not going to be able to determine or decide when we introduce it, it's just going to be there. 

But anyway, for now, we think it makes sense to try to delay it for a couple of weeks so that students know, okay, this is working with-, without AI, and now we're going to add AI on. But I would love to have everybody else's opinion on that too.

JANE: I think we just used the term IDE. So that's interactive development environment. So for anyone who's not a programmer or is not involved in programming education, it's like when we're, kind of learning to program our code and we're kind of typing into a system. We call that the interactive development environment.

IRENE: I wouldn't say in my context anyway that I would use AI tools to teach programming, but in fact, I don't even see myself as somebody who teaches programming. I'm there to kind of support them and how they can learn how to program. But when I'm teaching programming or being the teacher in the classroom, AI, it can play a part in all of this, but it's just one small part and it's a tool alongside everything else that we already use, that's still there. You know, the offline IDE, that's what I use at the start. 

You know, students are probably already using AI at home. You know, you hear this all the time. We can't control that. But what I can control is what's happening in my class. I really think as us teachers, it's our responsibility to create more opportunities for learning without AI. 

We need to not only spend those first few weeks building those basic programming skills and concepts and those ways that I've suggested, but also the time to build the relationships with the students, which brings in trust. Because when we do introduce AI tools, and only if we introduce them because we don't have to, that it has to be in that safe, trusting environments. You know, where students are going to be comfortable sharing what they're prompting, sharing what's coming back, and having those open discussions. 

But again, before we even go near AI, we have to teach them about AI. That doesn't have to involve any, any tools at all. So it was, it's refreshing to hear you say, Dan, you know, you're revisiting that the start of that book because that's our job, isn't it, to, to critique what we do, to reflect, to change things. And that's, that's what makes, you know, our jobs, I think so, so enjoyable in many ways that we can do that. But it has to begin with, you know, understanding what is AI, I mean, that, but that's different for every teacher, for everyone's comfort level.

JANE: Now, I completely agree. And just as a little reflection, I, I was talking to Leo, who works with Dan, and he was saying how in the second year he'd waited until lesson six. So until the students had learned the basic programming constructs and they could at least read some of the code before they anywhere went anywhere near to AI and then, Jedidah, I'm, I'm interested, from your perspective, do you think AI is being used by schools yet in computing lessons like, like you said, is this something more that students are using at home, maybe without their teachers even knowing?

JEDIDAH: My short answer would be no, it's not really used in computer lessons or in other subjects. But that being said, some computing teachers are more inclusive than others, so some can signpost you to tasks that they think, okay, you can use AI for this... other tasks, but then there's situations that are quite ironic where there's other teachers that ban the use of AI altogether. But you can kind of see how they're using AI themselves, and they're using it to maybe create worksheets or lesson materials. But that was, that was back when you could really tell when something was ChatGPT or not. But, yeah, these are some funny class jokes going on about, what chatbot did the teacher use this time? 

But, um, at home, I don't think it's really, it's not really behind the teacher's back, let's say, I'm pretty sure teachers, they do know what's going on, like I said. But I would say most students aren't using it to just be like, do all my homework for me. It's more they're asking the chatbot questions that they would normally ask if the teachers right next to them. So yeah, why is this syntax error coming up? Or why doesn't this line work? Rather than like give me five bits of code for question one, two, three, and four. 

And this is a concept I've been thinking on for quite a while now. So in 2023, late 2023, I pitched the concept of an AI shoulder buddy to Scottish Parliament at a Scottish Parliament conference. This is drawing on like my primary school experience. When you're in class, I don't know, you're learning how to tell the time or kind of money using pennies and everything. But you're stuck.

My primary school teacher would always say, if you're having trouble, first of all, think really, really, really, really hard yourself.

IRENE: Yeah.

JEDIDAH: Then if you can't think of the solution, ask your shoulder buddy. Ask the person next to you. And if you guys both don't know what you're doing, okay, then ask me and I'll come explain. 

So this concept I think applied to AI tools is similar to a coding companion where you're asking it very specific questions, but questions that you have to already have a grasp of the coding projects that you're doing first in order to ask the right questions. 

This same way that asking your shoulder buddy in primary school lessens the burden on an overstressed teacher where there's only one of them and 20, 30 odd students, it's speeding up the learning process where you're getting very quick, small questions answered quickly, and then moving on with the rest of your task.

JANE: I know, I understand what you think. What you're saying. There is kind of like this idea of not exactly peer programming, because it's not peer programming. It's just a little bit of help that you're going to get.

JEDIDAH: Yeah.

JANE: And I'm not aware of a term that we, we have the idea of an AI coding assistant, but this is more like a teaching assistant. Do we have a word for that kind of support? I, I've not come across one. Dan or Irene or Jedidah?

DAN: I've seen people capitalise the AI in the word pair programming. Have you all seen that? So it's like P and then capital AI and then lowercase r program.

JANE: It's a pAIr.

DAN: Yeah.

JANE: I don't know. I suppose... I agree completely with what Jedidah just said, that, you need to be able to read the code that you're going to be presented with in order to evaluate whether that's a good next step for you. 

And I do wonder, though, and this is kind of a different kind of, of thing because I'm interested in equity. In the study that I mentioned, well, what the teacher said to us was the output that came from the AI coding assistant for the students whose literacy level was high or who had experience already learning about programming. So they were pretty good at code reading. They did really well with the output from the AI, but the students with the lower literacy, they found it difficult. And so they were kind of almost saying that there was a an increase in digital divide rather than a reduction, which is what what we first got all excited about. So, Irene, do you think it makes it more equitable or less equitable?

IRENE: Less equitable. Absolutely. 

For the same reasons you have outlined, in my research, that's exactly what's coming back. The students, and we know who they are in the class, confident programmers, they were able to prompt. They were doing really well. 

The other students who are struggling from from what I've observed and from their, the data that was coming back, I mean, I was analysing so many ChatGPT conversations, student reflections, focus groups, you name it. But I was having nightmares about ChatGPT. But, um, yeah, it was it was clearly getting worse for them. So for me, it really, really widened the gap. Those students were getting extremely frustrated. The AI wasn't helping them at all.

But even more worryingly, on the surface it looks like those who are, who probably didn't need the AI help, right? They thought, oh, this is brilliant. So I'm getting all this functional working codes, but it was like, yes, there was no syntax errors, there was no red errors coming up. But actually when I was analysing the ChatGPT conversations, I could see errors everywhere. I was like, this code is not actually doing what it's supposed to be doing. And to me that was a worry because the misconceptions would build up, you know, even something as simple as indexing, right, it was indexing wrong when we were looking at strings and lists. So many examples of where ChatGPT was wrong. Now we know this. Okay, well, like we we know it can get it wrong. 

So yeah, I think it's, it's less equitable, Jane. I think, equity can only come if there's proper scaffolding in place. If the students, they have to be able to learn to read the code.

JANE: Yeah.

IRENE: I know, at a lot of universities, a lot of the research out there is pointing to tools with guardrails, and they're all like fantastic. But we don't have access to them in schools. And students are using ChatGPT because that's what's free and that's what's on their browser. 

So you have issues there between different schools, schools who are paying for maybe a tool that has better guardrails, compared to free tools. So, so I think it's, it's a huge issue. And look, I think until access is fair and consistent, I think the fairest thing that we can do as teachers is to ensure that our students can read code and that we provide those opportunities for them to learn as much without it, before introducing it.

JANE: You've just raised so many things in what you said. Some of our teachers seem to think that those with higher levels of literacy, with more experienced programming, they were okay, but maybe they weren't. Maybe they were getting almost a little bit lazy in terms of reading. I've heard about that or cognitive offloading, I think it's called. 

But then, as you said, there's also this thing of and the playing field is not fair because not everyone has access to the right quality of AI tools. And Dan, have you seen this as well in terms of with undergraduates?

DAN: Yeah, there's, there's so much to respond to. I guess, maybe if I could just add one more quick thing about the free versions, they have more privacy concerns as well. So typically when you're using a free model, if you read carefully what it's doing, it's, it's often using your data to train the models. And typically you have more control over your data if you're paying for it. So it's, it's definitely a huge amount of unfairness there because people who can't afford to pay for the models, their data is the data that's just, you know, freely made available to these companies. Right? 

It's great we have free educational access to tools like, GitHub Copilot, for example. I think schools need to do more. Schools need to do their part to make sure that all students have access to these tools. Just just like Irene was saying, we're at risk of making things worse if the students are using different tools. There's educational inequity there, but I do also want to report something that my colleague Leo Porter and I have been finding in our course. 

We're asking our students to complete these like large, open ended creative projects. And on these, we're not finding the gaps that we typically find on exams. So if you look at exam scores, students with prior experience, you know, they typically have like a letter grade better than the students who don't have prior experience or more sometimes. But on the on these projects, we're not seeing those gaps. 

And this is a really important outcome for me and Leo for this course. I think we really need to think about our assessments again, and what goals they're targeting and who is being treated unfairly by these assessments. I'm just really jazzed that we found a type of assessment that seems to make some progress on that.

JANE: The one thing I wonder when you talk about students creating projects, is that effectively a joint project that's part of the, part of the knowledge is actually come from the AI in terms of some of the components, in which case is it all just student work or is it student and AI work? And if the student were not to have the AI, would they be able to perform as well?

DAN: Yeah, I think without the AI they would not be able to complete these projects, or at least, the the scope of the projects would have to be smaller. But what's really interesting here, Jane, is somehow I'm okay with that, because one of our goals for the course is that students are able to create projects that are motivating and do something valuable for them or their families or communities. This isn't the only outcome we have. We want, like Irene said earlier, we want students to know how to work with code still, and we think that's very important, as well. 

But we're okay with the knowledge that students would not be able to complete these projects without AI. Because AI does increase what you can accomplish. I think Jedidah said it earlier, they pulled off a really cool hackathon project that they would not have been able to do before. 

So I think we just have to look at our outcomes for our courses honestly and say, okay, which of these do we need students to be able to do without AI? And for us, it's not all of them. Like, for example, could I do anything without my C compiler? Right? Not like… I don't remember how to do like assembly code or anything like that. I would be totally stuck. But I'm okay with that, right? I just, you know, I mean, time is limited, right? So different goals are going to require different tools. I would say.

JANE: Maybe there the the thing is, not only what you do by yourself but what can you do with either a partner or with, a computer to helping you, or with AI helping you? And then maybe it's less about independent coding and about being able to collaborate and construct knowledge jointly with an AI tool. Maybe. But that's kind of… 

Yeah, maybe that's really turning things on its head, and that's going to be exciting to look forward. So I think what I was going to, I don't know, is should we talk about research or should we talk about opportunities and risks?

IRENE: Well, I wouldn't mind telling a little bit about maybe my research.

JANE: Go for it. That would be lovely.

IRENE: So it's an in-depth qualitative study. But it was a design based research study. The students are at risk of becoming over-reliant on the tools.

DAN: Yes.

IRENE: They have gaps and misconceptions about AI. They don't understand how it works. They think it's intelligent. And they, they, they're overly trusting of its outputs. So they're not verifying what's coming back. There's a real need for ethical guidance because of this and... The use of the tools as well should be limited and monitored for, for all those reasons above. 

But most importantly, the students want, want this, okay? They they want conditional but limited access to it. They want to be allowed to experiment with it. But they, they, they want it in a way that, you know, the teacher is there, can see what they're doing and that they're open to all of that. As I alluded to earlier, it's really important that the trust builds in the classroom, because that allows us as a class to, to look at the outputs together. And the students came to the realisation themselves, oh, oh, the output is wrong there. You know, it was classroom discussion. They brainstormed. 

And the the big thing from the project really was about the co-creation aspect. So they would they evaluated their own conversations with ChatGPT. They were looking at their prompts. They were looking at the outputs, they could see the problems themselves. And they started to identify, well, what's a good prompt or what's a per prompts. So they co-created their own classroom guidelines. And it gave them a real sense of, of ownership and, and empowerment.

JANE: That's what I loved about your research was that you were involving students. There was student voice and Jedidah, I know that you've done research in this area as a student with student voice. So can you tell us about your research and what you found out about other students? How they were using these tools in terms of pros and cons?

JEDIDAH: So student voice was really the beating heart behind the research I did as part of the Young People's Advisory Group with Edinburgh Uni during my last year of school, and as part of a team of eight young people, and the professors at Edinburgh, we designed research questions to ask other secondary school pupils. So a lot of students. 

We also noted what Dan was saying about privacy issues. Definitely in my focus group there was awareness of the ethical, environmental issues, that AI models pose, as well as the willingness to kind of try learning in almost a traditional way first. And then if AI is needed it, then it's there.

DAN: I feel like one of the most important skills, like it's always been important, but maybe even more now, is self-regulation, right? Like the students have a way to generate something. It might not be a good submission or whatever, but. But sometimes even I find myself using the AI too early. In my head, I'll be like, what's come on? Like, we're not going to learn like this, what are you doing? 

But I'll still I'll still do it because it's right there. It's like, I don't know. It's like junk food. It's like chocolate. You know? It's like there's a reason I don't have any chocolate on my desk, because I would, you know, I mean, it would just be gone, right? 

So self-regulation, I think, is something we can help instill in students, especially when working with AI and understanding the trade offs of, like, Jedidah just said, yeah, you can get it to produce your report for you at the last minute, but maybe that says something more broadly about maybe your time management or how many projects you're taking on. Maybe you're taking on too many things. Right? 

There's there's just there's so, so much metacognitive awareness and self-control that's involved in working with AI that I, I don't know if we talk about that enough.

JEDIDAH: Yeah, I kind of had that. I would call it a moral dilemma this week when I was trying to get my uni coding work done. It's very different to school coding, where we've got weekly tasks that we needed to make. Based on the lectures that week. And I was finding this the one of the last exercises in this worksheet, incredibly hard. 

So I was there, my brain was just scrambling and I, I purposely did not want to use ChatGPT or Claude or Gemini to just give me the answer because I was like, at the end of the day, I really do want to know what's going on here. So when it comes to me personally, I, I, I'm thinking, what is the trade off? It's either do I get a speedy output and a high mark on my assignment or I sit with a little bit, try and problem solve because I think it's a very key skill and then I don't want to say risk, but, you know, leave chance for me not getting as good of a mark on this particular assignment, but at least I've come away with a bit more understanding of the coding concept.

JANE: And it is that thing of, are you at university just to get the assignments done? Are you kind of like a worker who's just producing output? Are you there to learn? 

It's such a temptation, though, isn't it, just to get it done. But actually the point of being there is, you know, is as Dan said we need to learn rather than just get things done. But there's, there is such pressure. 

Irene, you're just writing up your PhD in computer science education. What further research would you like to to explore you personally or see done in this area once you're finished writing up?

IRENE: Well, okay. I mean, it'd be no surprise, but I'd love to see more research happening at the second level. I just think we need more classroom based studies, getting teachers involved. 

Much of the current research is focusing on the tools, whether that's involves benchmarking them or developing tools with guardrails. They're important, but there's just less research on the actual effects of these tools on the, on the learning. I know, I know that's changing. That's a, that's a positive thing. 

But from a kind of a bigger picture point of view. Again, I'm speaking as a teacher and I'm not just speaking as a computer science teacher. I'm a mathematics teacher, but fundamentally, I'm a teacher. My job is to prepare these students for, for life in not just a software engineering role or as a computer scientist, but I'd love to see more work done on AI literacy, you know, across all subjects. I think as teachers in areas like history, humanities, languages, you know, they find this very, very intimidating. 

And I'm doing a bit of work on that in my own school trying to support the, the other teachers in the school, and I know there's great work happening in Scotland. I saw I see the stuff you're doing Jedidah. I saw the, the Scottish presentation at UKICER I know the Raspberry Pi Foundation have fabulous videos, you know, that are accessible to all. But I, I think we are changing. It is changing in Ireland, but kind of up until this point, most of the CPD for teachers out there, it was all on the tools, you know, how can we use the tools, the tools, the tools, but not enough about the learning. So I'd just like to see more of a focus on this.

JANE: So I think if I precis that, is it, it's kind of let's find out what effect it has on the, on the, on the young people and what is the right pedagogy. What is the right instructional approach?

IRENE: Yeah.

DAN: Speaking of future research possibilities, like if, you know, if somehow we all of a sudden have lots of money or something, the there's some research on slow AI.

JANE: Yeah.

DAN: Because Irene, in yours, your students don't have the AI integrated into the programming like they have.

IRENE: Yeah. And the reason why we use ChatGPT, Dan, is because that was the only tool when I started the study that had guidance for 12- to 18-year-olds. So they are learning Python and it was an offline IDE. But because it did come up like from an ethics point of view, it actually was the only tool, that ethically I could get approval for because at the time it was the only tool that had any sort of guidance.

DAN: You know what, though? I feel like it might be the way to go because you're I mean, ethically, you had no choice. But it did introduce a layer between students and code generation. Like there was some friction there they had to engage with the code from ChatGPT first. Right. And there's other research, where they're doing that. 

There's, they're intentionally slowing students down. Not too much, because if you slow them down too much, they're going to say, well, like, get lost, I'm just going to use my own tools now. But if you slow them down a little bit, maybe it will still be palatable for students. And you can introduce more learning. If, if you can slow students down enough to engage more with what's coming out of the AI, right.

JANE: Now, that exactly. I think that's absolutely spot on, Dan. And I think we need to do that research because that slowing down, that friction, that clear, it's almost then becomes intentional because, you know, it's a different part of your learning. And then you can do what Irene suggested, which is then reflect on what what that means and what that how that impacts your learning. 

But we really are coming to the end now. So I've got and I'm going to have to ask for quick answers so that I know it's going to be hard. So to wrap up, I'd love to hear one piece of advice for the educators who are listening to this and the students as well, who maybe are listening to this, around AI in classrooms, using AI in classrooms to teach programming. And I'm going to start with Jedidah.

JEDIDAH: So to my fellow students out there, I would say that AI tools, they're not a crutch. They're more of a springboard for you, you know, just think back to primary school. If your shoulder buddy did all your work for you when you get to the test, even a primary three tests, you wouldn't go very far. 

So now that you've got access to these tools, you know, take pride in your own work and your integrity, but know where and how to access help if necessary.

JANE: That is just wonderful. Thank you so much. Irene, what about you?

IRENE: I'd say avoid the hype that's out there. I'd say, you know, create plenty of opportunities in the classroom to learn without AI. In fact, probably even create opportunities to learn without technology. Okay. Because there's so much AI built into them. 

And if you do decide you're going to experiment with AI, then teach your students, teach yourself, actually learn yourself a little bit about AI first. There's plenty of resources out there. Teach your students a little bit about AI. And then finally, if you're going to introduce it, then, you know, create the opportunities for your students to experiment, but get them to come up with with guidelines, you know, discuss what works, what didn't work, what might be misleading. You know, most importantly, teach them to be be critical.

JANE: Be critical. Dan. Final word.

DAN: Maybe just remember that we're on the same team as students. They don't want to cheat. They don't want to have low achievement. They want to learn. That's the mindset we have to have. 

I've seen so much discourse about they are, they're cheating, they're using AI for everything. And it's like, maybe they're doing that because your courses are not, are not working in the AI era. Right. 

So so just let let's try not to blame students for you know, for them using AI. Let's really look at our courses. I would love to encourage people to just just do something. So I've tried things in my classes and have blown up on me for AI, and the students never react negatively to it. In fact, I've gotten some of the highest course evaluations from students with disasters. Like I've tried assignments that have just not worked out, and it's like, okay, five out of five. It's like, because students know like students know that you're putting yourself out there and trying stuff, right? Like they're not going to they're not going to come at you if it if it goes wrong, they probably expect it to go. They don't know what's happening right now. Nobody knows. Right.

JANE: So I think that's I absolutely, completely agree. And this has been it's been just so illuminating and just so interesting. And and I feel like I've brought together these, three amazing researchers who've been able to look at this really complex and gnarly topic. 

And we've talked about reshaping authorships, thinking about student agency, AI literacy. But most of all, including our students in this journey with us and thinking how we might change what we teach and how we teach. And I just think...I...I, I just feel such privilege to have spoken to you all. Right. So thank you so much for joining me. And, I know I'm going to carry on talking to you all, so, that was brilliant. Thank you.

IRENE: Thanks very much, Jane.

DAN: Thanks, Jane.

JEDIDAH: Thank you, Jane.

JAMES: Thank you so much, Dan, Irene, Jedidah, and Jane for that brilliant conversation. 

As I said at the start, we'd love to hear your thoughts. Do you agree that we should ditch the term 'vibe coding' in the context of programming education? What do you think of the use of AI tools to support students learning to program? Let us know by emailing us at podcast@helloworld.cc. 

Next week we're bringing you a great discussion around the pros and cons of using Scratch, Python, Java, and other programming languages to teach coding skills at different learning stages. 

A reminder. You can find the most recent issue of the Hello World magazine over at HelloWorld.cc where you can subscribe and read the digital version for free. Thanks so much for watching. 

See you next time! Goodbye.