(cheerful music) - I used to be a little reluctant about AI, but we cannot fight against it because you don't stop evolution. - Every big company, every organization now is using it and so we can't really avoid it. - I really need guidance from experts on using GenAI for education. - Hello, welcome to our latest episode of the "Brighter Thinking Pod" from the International Education group of Cambridge University Press and Assessment. Today, I'm your host.
My name is Laura Kahwati, and I'm an education futures manager at Cambridge. We created our "Brighter Thinking Pod" to support teachers and school leaders around the world. Each episode brings you helpful advice and interesting conversation from authors, teachers, and academics. As you may have guessed from the teacher voices we heard in the intro, today we're going to be talking about artificial intelligence, specifically, teachers changing perceptions of the technology.
Remember, all the links and info that we discuss today are available in the show notes for your ease. (cheerful music) Let's meet our guests. Today I'm joined by Rachel Logan, school liaison and insight manager, and Andrew Field, senior education manager, both from the Insight, Ideation and Impact team here at Cambridge. So welcome both. - Hello. Hello, thank you very much for having us here. - Yeah, thank you. Great to be here and excited what we can about to talk about.
- Oh great, well it's lovely to have you both here as well. I'm hoping that you won't mind if I start by asking you for a definition. So could you define for me generative AI? - Over to Andrew with this one. - Well, we've had loads of explorations with this and we know now that people were initially inspired and enthused about AI in late 2022 into '23.
And generative AI, as we're all emerging, all understand now is far more focused about specific area of AI rather than the wider aspects about machine learning. Generative AI, as we're seeing it emerging within teaching and learning, is all about using AI creatively to throw educational challenges, problems, exciting and in interesting areas to then see what can be generated to support us.
So in terms of defining generative AI, it's an interesting one because any definition we give you right now will change in the next minute. So I'd far rather focus it much more about what we're looking at at Cambridge, exploring things like the educational impact of these things.
And from the history of technology and the impact of items like this, we don't want generative AI to be seen as one of these historical tech things that get thrown at teachers to explore and then see what value might it be. Actually here, teachers need to be rigorously analyzing all of these tech promises to see what's the difference is it really gonna make for my classroom. - Oh, great. Okay, so there's actually, there's a lot of nuance there.
And there's also lots of little pockets of expertise that we want to look into, including, like you say, teachers in the classroom. Rachel, is there anything you want to add to that definition? - No, that sounds good. Sounds good to me. - Oh, fantastic. So we go from thinking about a definition of generative AI to actually thinking about how people respond to this.
So from your research with our Cambridge teacher community, would you say that there's more excitement about generative AI at the moment or more skepticism? - Well, I've got some great stats to tell you about that. So last year when we ran a big survey with teachers, a lot of them felt very anxious about what generative AI meant in education and what it meant to them. And we were really curious to see if their attitudes had changed over the last year.
So we just ran a really great activity with our panel of Cambridge teachers and I also used some AI to do some sentiment analysis on it. And I was really impressed with the results. So 27% of our responses were optimistic, which I thought was really encouraging compared to where they were last year. But also the AI analysis said that they were determined, hopeful, appreciative, confident, accomplished, but also that they were doubtful, frustrated, and curious in different measures.
And I thought that was a really nice way of picking up that, like you said, it's nuanced. They're not just all optimistic. They want to know what it means for them. What can they do with it? What does, you know, what does it really mean? And they're still not kind of fully confident. You know, people didn't rush into that and say, "Yeah, I've got this. I know what I'm doing." So that was really interesting.
- And so linked to that, it really does come back to sort of perhaps grandiose terms that we might throw around sometimes about what's the consequential validity of AI as we move towards the 2030s. But actually I want teachers really to be saying, "So what?" If someone says to them, "AI is gonna be fantastic, it's gonna transform the world." They actually need to be able to say, "So what?" In a polite and positive and enthusiastic way.
But what we've really got to explore here is why generative AI specifically is fundamentally different to all of the previous technologies that have been promised to teachers. It is really, really easy to be cynical, and that's the lovely aspect I've seen about all of the great research Rachel and the wider team have been involved in, where we've actually seen teachers saying, "What can this do for me? How can it help my learners?
How can it help me with my planning?" And beyond that, what's the actual transformative, impactful difference of generative AI? It isn't just a thing that's gonna make our lives easier because we can organize our data better or create a lesson plan faster. This is actually about things we've never even been able to consider about large data models, exploring and developing solutions that way, analyzing data in ways that previously weren't possible. And that's where it's exciting.
If you considered where things like the internet was way before many of you were even born, we're talking sort of late '90s, let's go for that then. If you consider some of those things where people first saw Google, teachers were concerned about it, and from an absolute source of validity, they felt that learners would be typing in questions, they'd get the answer straight off Google, and then homework wouldn't be able to exist anymore.
Now there's obviously discussions about homework and its importance and its practical reality, but it's all really focused on the skill and attitude and approaches and general brilliance of teachers around the world. Teachers love generative AI when it can make their lessons more interesting, more attention-grabbing. There's loads we can do with it, and what it's absolutely vital we do and we're gonna explore in some of the discussions today is actually, what is the difference it can make?
And also, how can I best make use of it in my class? Or when do I actually open it up to my own learners to say, you choose, do you want to use AI? Do you not want to use AI? And take us on that journey 'cause it's the journey that is the most important there about where we can see why did you choose to use AI there or why when you used AI to source your materials or your writing and it changed what you were about to write, was that a good thing or actually a negative influence?
Some of the experiments you find that it may turn it into much more effective and long words or longer paragraphs, but does that actually get the job done that you're trying to achieve? So it's trying to light the fires within all these colleagues to explore how can generative AI really make a difference. - Absolutely, and I think there's a couple of things I want to pick up on there because like you said, it's really about how we use it in terms of making a difference.
And that's why I really like the way that Rachel, you framed that research that's been put out in terms of all those lovely emotion, adjectives used, whether the teachers feel hopeful or anxious or confident. And I think that's really important that that human element added to the kind of preconceptions that Andrew may have been talking about. But the other thing that I think is really great is that so what question, because you're absolutely right.
Like anything when it comes to education or teaching and learning, we want to be able to say, "Well, what does it look like?" And I really love the fact that what that seems to come down to for you is empowering teachers to actually be able to do something really successful in the classroom. If you don't mind, I just wanna ask you a follow up question there. How do we empower students not to feel like it's something that they might want to just say, "Oh, I'll just put it into AI."
- For me, it comes down to the skill of the teacher. If you look at it at its very most basic level, generative AI could be considered merely a tool. So just as your classroom previously was filled with fantastic books, great resources, opportunities for reading nooks and different spaces for creative writing and art and anything else you wanted to do, you could see generative AI as part of that toolbox.
What it really has got to be is the focus on my absolute favorite thing that Cambridge has been involved in, the Cambridge Learner Attributes, Those absolutely brilliant attributes where we're focusing on encouraging learners to be creative, helping learners really be inquisitive to explore what they could do. So just like a great teacher doesn't tell learners what to do, they inspire them to explore and see what might they do.
We want to enable teachers to be confident enough to open those doors for those learners to see, "Okay, here's a challenge for you. What might you do?" And it's not necessarily even just saying take it or leave it, it's doing things like, take your current question, put it into AI, and then us together, we're gonna explore the output, critically evaluate that.
Because where some of the areas where there is genuine concern is, are we offsetting these great critical evaluation analysis skills to AI? And what we should be doing is looking to explore some of the more mundane tasks and getting AI to take care of those. So the learners themselves can be exploring the creative elements, the adding the value. But just as people used Clippy years ago in Microsoft Word, that's where AI can be used here. It's inspiring people to see what might be possible.
- I want to jump in on that as well about the point about critical thinking and ownership as well. And I think like students really want to feel proud of their work and really connected and engaged with it. And you can use AI as a critical friend to help them build those kind of critical thinking skills. Like, is this good enough? What else would you do? Is this accurate? And they kind of want to catch it out as well.
They want to prove that they're better than it, a bit like, you know, playing a chess computer and you want to get it, you know, make sure that you win. And I think that, you know, I've yet to meet a student that didn't want to feel proud of their work and kind of, yeah. And it's that thing as well, like, if they're gonna use it, they're just going to... If they're gonna use it to like cheat, they're gonna be self-penalizing.
And it's learning how to use it, it's learning to, maybe you're not that great at setting up a writing frame, but you could use AI to give you a starting point rather than a blank piece of paper and then go in and, you know, make it your own, change words around, back up your sources. If you haven't read the book, it's not gonna help you. It's kind of like making sure that you're using it as a tool and that ultimately your student voice is the important thing that's coming through here.
- But what a great point about reflective practice, the absolute best way to develop your expertise, exploring ways to, better next steps, your formative development is to reflect effectively. Now the AI tutors that we're seeing emerging, I'm far more interested in actually learners themselves choosing to, let me put that to this AI assistant, this Copilot, to explore what is of value.
So when they are drafting something, actually we need to be exploring the effective ways where you can use Copilot to then reflect back and consider. The aspect with a lot of the commercial offers at the moment though is they're very good at saying, "Yes, of course I'll do that," and then maybe not quite doing it. Now some might say that's a limitation of the current offer. I would also say we've got to make sure that the educational impact of all of these things comes at the forefront.
It's so important that a learner is able themselves to critically evaluate the whole process. Does it take longer to ask the questions of AI than it would've been just for them to answer the question themselves. Now that will change as technology and other things get embedded and we can explore further. But it fundamentally does come down to the Cambridge Learner Attributes, helping learners to explore and inspire to see what amazing things they can do.
That will be significantly empowered and enhanced with AI. - Absolutely. And I have heard you use the phrase before, Andrew, that it is always about the teacher and not the technology. And I always like to think of the Cambridge Learner Attributes as something that teachers hold within them as well as students.
So what I'm getting from both of you here today is that when it comes to the use of AI, it is about encouraging teachers to use good judgment about how and when it's best used, but also encouraging students to use good judgment and in terms of how and when it's best used. - Definitely, definitely. And using it to make space for the really important conversations. So you know, if you're spending a lot of time copying big chunks of information from one document to another, that could be AI.
And then that means that your time in the classroom talking to people, the students that you know really well, that's where the value is, isn't it? So using it to do your boring tasks. - A lot of what we do, we're always thinking about busy teachers and the amount of work that they have on their hands. So I think many would be excited about the idea of ways in which they can use AI to actually free up their time for their own creativity and thinking.
- Just as long as that freed up time then isn't identified by a management individual in their school to say now to do some more admin. It's really important that we continue these quite rigorous conversations about best use of teacher time. The teachers, if they're given time to explore, can do amazing things. We're not about adding more bureaucracy for teachers.
We're about using AI, as Rachel and others have mentioned, about actually getting those more mundane tasks done faster so you can focus on having a creative space. AI is gonna do amazing things that we're not even sure of yet, but it's not about these images of robots taking over the world and things. This is about teachers being empowered even more than ever to do inspirational things for their learners.
- Now speaking of AI doing amazing things and teachers doing an amazing and inspirational things, can I ask you both, what in your experience are teachers most excited about? Where do you think AI can provide value to teaching and learning? - I've got some lovely quotes actually about that.
So one of our teachers said that they thought that an AI language assistant would be amazing that a conversational language tool that you could speak to and it would speak back to you and it would give you tips on your pronunciation and that kind of thing. And I think when we're talking about our international schools, we're often talking about bilingual learners as well.
So something like that, that takes the, perhaps the pressure off the teacher having to be the person giving the feedback all the time, but also the students might feel more confident to practice with an AI tool. They maybe feel not judged in any way, or you know, more confident to talk to it in a more natural way. So that's kind of, that was a suggestion from my teacher about what kind of tool they'd like to see in the future.
- I think what we're exploring here is how teachers can empower their current practice even further. That's kind of the first level of adoption of AI and it's also being willing to explore and try things out. Like some of the best activities I remember from way back when I was at school was when you were set these challenges without actually being given the tools to use it. The part of the work was to explore what was the best tool to evaluate where the best use was.
Was it worth using a computer in those days to program something when there's some other way of doing it? I think that's where this Copilot, this co-intelligence of AI can really assist learners. And it's not about presenting problems and then saying, you've got time to solve this now. It's actually about exploring how we can use AI with real impact in the classroom. Helping learners to say, "I did use AI, and here's the difference it made and here's why I would use it in this way in the future.
But also having tried this, why I wouldn't use it." I think that the items we've explored already today where it's about this coaching opportunity, this ability to throw ideas out there and get some suggestions back.
the current AI chat bots and assistants are really anxious to please, which is good, but if I throw out an idea which is maybe slightly contentious, you want an AI coach to really come back to you and say, "Well that's not a bad idea," with a praise sandwich and, "I like that suggestion, but perhaps you might consider these ideas too." Helping learners reflect in that way is a really impactful way of doing things.
If you think about this at scale, with a large class of learners, each of them could have their own interactions with the tools that they can then use. And then it leads into far more effective discussion in smaller groups than whole groups. So it's about not necessarily buying or investing hugely in expensive or massively costly additional tools. It's about encouraging those companies selling those offers to actually embed them into existing offers.
A lot of things we've been exploring at Cambridge is about adding value to what we've already got. The Cambridge Content is amazing. We're all aware of the enormous history there, almost the quality, the absolute research-based items. Now imagine these things being able to present themselves on demand. You've got a request, you've got a suggestion, but if you're logged in with Cambridge, then you can get the suggestions from the Cambridge Content.
There's aspects there which we're still exploring, but that's where a lot of the research we've done with teachers has been, you know, I know I can use ChatGPT, but where's the Cambridge Content there? How do I know I can trust it? I used to say when I was teaching that Wikipedia was the most trustworthy website ever because no one trusts it, which means you always double check your sources. Oh, you got that from Wikipedia, did you? Well, you must make sure you check that twice further.
And it's interesting and innovative ways of using current AI tools that we need to explore. It's about always asking those questions, how do I know it's right? Where's the source material come from? Who's paid for this? Where's it really come from? And those are not existential questions of 2024, they're humanities questions from 500 years ago. So it's about bringing all of this together. - All of that sounds really exciting to me.
And also in terms of thinking about it when I hear you talk about it from being a former teacher myself, but can I ask you what kind of fears there are around generative AI for education and also how well-founded those fears are? - Well, I think if we come back to the skepticism as well, like teachers, you know, we've already talked about 'em not having a lot of time, this is just another thing for them to get to grips with.
They've got to keep on top of, you know, keep a step ahead of the students and they're not really convinced it's actually gonna be all that great in the end.
And I was just thinking about the kind of level of skills that we expect from our teachers now because they've got, while Andrew was talking about the things to do with generative AI and the types of tools, I was thinking we were expecting these people to have really strong data visualization skills, which, you know, they're not leaving school without. They're not getting that from teacher training.
Why do we think that they're gonna walk into the classroom amazingly brilliant at data analysis and reporting and tracking and being able to give feedback to students? Like that's where AI can help them. And once they start to see where it can help them augment what they're doing, do things even better, do things quickly, then they start to shift towards thinking, oh, actually this could be really useful.
And the skepticism, like now they're saying, "All right, I know that it's not gonna replace me. Like it's, you know, it's useful but I'm still the most important thing. And if everything in the classroom, the students come to me first. They trust my judgment over a piece of generative AI. They trust me more than ChatGPT, therefore I still know my value.
So I think it's kind of like skepticism, anxiety, and then a kind of a quiet confidence that they are still the human in the loop and the most important person in the education system. So yeah. - And Cambridge has a real role there in the international community to reinforce the importance of teachers. We did so in May 2023 when we published the thoughts and ideas about the impact of generative AI.
And actually even though our research and obviously our findings and discoveries continues, it still is fundamentally focused on supporting that teacher. There are gonna be amazing things with generative AI and if you'd flown back 10 years ago exploring different opportunities with technology, it was emerging then. Technology types always promise that it's gonna be a better tomorrow with all these great opportunities.
Actually it's the teachers who create that conditions for that better tomorrow, the teachers who then inspire their learners. So I'm nervous to just say that technology, that AI is another tool because it's clearly not. It is transformative in its potential, but it's the teachers who are gonna make that happen. So we're not here about selling technical products and services. What we're selling here is high quality Cambridge education with real impact for now and, well, for the continuing years.
- I think that sounds really exciting the way that you put it like that. And it makes me come back to Rachel's wonderful phrase of the human in the loop being the thing that is the most valued all of the time. And speaking of the human in the loop, whether that be a student or a teacher, can I ask you, what is the most innovative or impressive use of AI that you have seen as an example? - Can I respond with a different question to start with though, slightly differently from that? - Absolutely.
- It's in terms of AI experts. We're not AI experts and we were never pretending to be, and I would be so bold as to say anyone who says they're an AI expert actually isn't because it's such a fast-moving and transformative time at the moment that singling out items for, that's the best use of AI is actually putting the technology before the learning. So what we've perhaps gotta do with this, the most impressive use of AI that we've seen has hardly got anything to do with AI at all.
It's where there's been really good structures, really good setup, really good creative opportunities that perhaps have been empowered by AI. But actually when it comes down to it, AI has been the enabler, the tool and the opportunity that's allowed a teacher to do more amazing things. For me, Rachel and I were involved, there was a discussion with two brilliant teachers from our Cambridge community where they had different levels of confidence.
Actually they were both brilliant, but one felt that they had hardly any skills with AI and kept talking about all of these cognitive load and the skills and the discussion opportunities and all these things they were doing in their classroom. And we were thinking, this is brilliant. They're doing all of these things, but they weren't even able to step back and realize that AI had enabled them to do those things.
The other colleague was fully aware of the impact of AI and was doing brilliant suggestions and ideas in a mathematical focus. But again, it's about using AI to get things done to open up these further opportunities. So I think the question is perhaps a challenging one for us in that I don't think you should or we should ever separate what's a great use of AI. It's actually where's the educational impact in what we're doing?
- And I would agree actually in terms of thinking about teaching and learning in general. Sometimes when we use those superlatives like most and best, we're actually missing out on considering what might be effective, what might be good for one person, what might be good for another person. Is there anything you'd like to add, Rachel? - I just was thinking about research and that kind of methodology and that it's often we measure things that are easy to measure.
And I think that AI, you know, improving your report writing, your emails is an easy to measure thing, isn't it? And that's where we might end up saying, "Oh, this was really good at this." And not saying, "Oh, the impact then was that I didn't feel as stressed walking into this situation. I was more confident. I thought more about the language that I used when I was questioning my students."
Like that's not as easy to measure, but they're kind of, you know, this tool saved me 15 minutes, therefore it costs me, cost benefit to my business is this. You know, let's kind of think about just what you've kind of saying that the good ones, you won't even know that it's there. You won't even know it's got AI in it 'cause it'll just be so good, it'll just work. You know, let's just make all of our stuff work, like, number one.
- And with that, one of the best pieces of feedback we got a number of years ago was this brilliant teacher in Porto who said, "If you could save me five minutes a day, that's a golden amount of time for me over a week, over a month and so on." And it's that view that I'm really focusing on, helping the existing materials that we have and also developing new products and services, which may be powered by AI.
But actually, and we've done loads of research on this, teachers don't really care about knowledge graphs or the LLM or the RAG analysis. They care about saving five minutes a day because they've been able to log in effectively, find data they need, find the support materials, find the textbook items, find the resources they need, and it's that which is the empowerment enriching item. And no one ever phones up IT support to say, "Just wanna let you know it's been a fabulous day today.
Thank you for your efforts." That's where we're gonna see the real transformative nature of AI where it just has enabled these things to happen. - And like you say, that five minutes a day, doing it in little ways here and there and often and building up. Now my last question is arguably quite a controversial one and a big question and definitely hard to measure.
So do you think that we are, and I put in quote marks here, breeding a generation of well-informed idiots as one teacher put it, and what do we mean by this and what are the consequences? - So we loved it when that teacher put that rhetorical question to us, because obviously that isn't the situation we want to be in.
And what we want to be breeding is a generation of students that are really confident to articulate what they want, to be able to express themselves, to be able to question things, to use those higher order critical thinking skills, you know, to back up their sources, to kind of, and we talked already about owning a journey, owning the fact that it's your own work. So definitely Cambridge is definitely not doing that.
We're definitely making sure that our students are ready for the world and can use the different tools in the right way that kind of come across their path. And then, you know, we're also, our Cambridge students are the teachers of the future as well. So we want to make sure that they're well-informed, well-grounded, articulate.
- And that challenge of that question is a really good point because if you look at it in a simplistic way, you could say, well, now I can ChatGPT anything on my phone and appear like a complete genius. But as we know, that is useless without the ability to talk about things further, the ability to analyze your thinking that the critical thinking skills that we've been talking about, it's about empowering our learners with all the items that we've been repeatedly mentioning.
But the Cambridge approaches to teaching and learning is about challenging things. Why has someone said that? What might their viewpoint be and what might the considerations be? And there's a danger, and this is this fear about AI that it's gonna take away all of our creativity. And what we need to do is actually use AI to enable even more creativity, helping organize data sets more effectively. All of these potentially quite boring technology items about information architecture.
But imagine you have everything available to you, what do you then do? What you actually do is you still do your traditional teaching and learning, but then you can inspire and explore new ways of doing things. Spoke a lot about VR and things like that previously where you pop on a headset and then your learners in your class could be exploring anywhere in the world.
But someone walking past that classroom would just see a load of people with looking like sort of funky headsets on just moving their head around, and like, it's all about how we actually use this technology for impact. And it's actually the complete opposite of what was posed by that questionnaire, generation of well-informed idiots. It's actually a generation of inspired, enthusiastic people who are gonna do amazing things in the next century. And that's the area that we're really focusing on.
So no way is it a generation of well-informed idiots. What we're talking about here is a generation of learners who've been inspired by the Cambridge Learner Attributes and actually willing to make a significant difference to the world. And we're talking the climate impact, we're talking the impact they can have in their local community, the global community, and AI is a part of that, but none of that happens without that well-rounded, experienced, highly trained Cambridge teacher.
- I think that's a wonderfully reassuring answer to the creatively posed question that we were given there by that teacher who took part. Thank you both. This has been really insightful and it's been wonderful to hear you both talk about how we can empower teachers and students to feel really good about generative AI in the classroom. At the heart of everything you have said, you have talked about the value of the teachers and the value of the students.
And if that means that they can be helped with generative AI and empowered by it, then that's a wonderful thing. That's all we have time for today. Thank you to Rachel and Andrew for being such fantastic guests and sharing some really useful insights. Don't forget to tell your friends and colleagues about us and rate our show on whatever platform you're listening on. Our show notes have lots of useful links that we've discussed throughout this episode, so be sure to take a look at them.
You can also follow us on X and Instagram @CambridgeInt. We're going to leave you with some more of those fantastic quotes from teachers around the world. Thanks for listening. We hope you join us again soon. - I used to be a little reluctant about AI, but we cannot fight against it because you don't stop evolution. As educators, we have to learn how to deal with AI and how we can use it wisely when teaching learners.
- Every big company, every organization now is using it and so we can't really avoid it. And what we're doing is just to support this generation. In fact, we had an insight on how to use AI in our teaching and learning. So it is something that we need to accept, it is with us. We just need to see how we can support them and lead them or guide them to use it correctly in their teaching and learning.
- Teachers are also using AI to save time on tasks by grading so that we can focus more on students. Overall, AI is becoming a helpful tool in our classrooms. - Obviously AI is not gonna replace teachers, but then the best thing as teachers, as educators, is to just get this generation to use it wisely. - Using AI responsibly and with the right frame of mind to develop learning will enable us to produce the best lessons.
Not to let AI take over, but to work alongside the emotional intelligence that we as a human brain and as teachers have. - GenAI is really new to me. I only know a little bit about GenAI, so I really need guidance from experts on using GenAI for education, preferably if they are science-related and also education-related. (cheerful music)