Creativity in the Classroom with AI Technology: a conversation with Joanne Taylor about PATT40

Prepare to have your perspective on art and design education challenged as I sit down with Joanne Taylor, after the PATT 40 conference. She's here to share her thoughts on research from Oslo Metropolitan University on how AI text-to-image generators like Mid-journey, Dali-E, and Stable Diffusion are revolutionising the field. We dissect the conference highlights and reflect on the importance of staying abreast with academic research, you'll discover the keys to nurturing evidence-based learning in our ever-evolving educational landscape.

This week's discussion goes beyond mere tools, probing into the ethical terrain where technology meets education. We're tackling topics like combating bias, fostering originality, and ensuring that our students' innovation is expressed with clarity and depth. With Joanne's take on the ethical implications of AI in design and the importance of traditional skills in harmonising with new technologies, we share some thoughts about how AI can be used to guide your students into the future.

(Text generated by AI, edited by Alison Hardy)

Ringvold, T. A., Strand, I., Haakonsen, P., & Saasen Strand, K. (2023). AI Text-to-Image Generation in Art and Design Teacher Education: A Creative Tool or a Hindrance to Future Creativity?. The 40th International Pupils’ Attitudes Towards Technology Conference Proceedings 2023, 1(October). Retrieved from https://openjournals.ljmu.ac.uk/PATT40/article/view/1350

Links
Dall-E
Stable Diffusion
Midjourney
Paul Russell at Loughborough University

To  connect with Joanne:
LinkedIn

 jtdesignandtechnology.com

Ciaran Ellis posted a thought-provoking question on LinkedIn recently: Do design decisions involve value judgements?

What do you think? Join the conversation over on LinkedIn and let us know what you think. 

Transcript

Alison Hardy: 1:03

You're listening to the Talking D&T podcast. I'm Dr Alison Hardy, a writer, researcher and advocate of design and technology education. In each episode I share views, news and opinions about D&T. So welcome to this week's episode of the Talking D&T podcast. This is part of the PACT 40 series, where I've asked a range of people who were involved in PACT 40, either attending or contributing or organising to come and share some of their experiences from the conference, and this week I'm with Joanne Taylor, who attended the conference. We had hoped to meet at Liverpool, but circumstances meant that I didn't actually make it, so this is really our first time talking about the conference. It's great to have Joanne here. Joanne's going to share some of her thoughts about one of the papers, which we'll come on to in a moment. But first of all, joanne, would you like to introduce yourself, say who you are, where you are and what you do.

Joanne Taylor: 2:02

Thank you. Well, firstly, thank you for inviting me onto the podcast. It's a little bit overwhelming. It feels a bit like being on desert island discs for design and technology teachers. I'm Joanne Taylor. I'm a design and technology teacher, mentor, trainer, now consultant and researcher. I studied at BEd at Nottingham Trent in the 90s, started teaching in 99, completed my master's in my early career and now I'm going back to study in with a PhD, which is what led me to contact you last year for some advice on how to get that ball rolling.

Alison Hardy: 2:41

Yeah, so that's a warning to people who contact me who want to have a conversation. This is where you end up, right? You end up on the podcast and advised very strongly to attend conferences. Well, that's what.

Joanne Taylor: 2:52

I was about to say I was really grateful not only for your time but also the suggestion that I attended Pat 40, which, I'm a bit embarrassed to say, I'd never heard of before, despite many years of teaching D&T.

Alison Hardy: 3:05

I think it's just one of those things that sales under the radar. It kind of. You know, it is mainly researchers who are in working in academia and working higher education that present. We made a concerted effort this year. More teachers were involved, but it is quite difficult because it always falls during the academic year. People are teaching, so it is quite tough, but no, it's good that you weren't, so you got a lot out of it. We kind of caught up on the last day when I attended, virtually didn't we and I know you were absolutely buzzing. So three months later, are you still buzzing?

Joanne Taylor: 3:36

Yeah, absolutely. It was fascinating and really welcoming. It's a bit daunting going into a new space with people that you don't know. My expectations were just I was nervous, I was excited. I was really interested to hear what's going on and I learned that it's a really vibrant community and everybody's welcome.

Alison Hardy: 4:08

Yeah, it is. The PAK 40 conferences are unknown for. Yeah, vibrance, that's a really good word, but also been really welcoming. So, yeah, so do you want to start off then, joanne, by telling us which paper you've picked? Well, it was really difficult to choose.

Joanne Taylor: 4:24

There's over 70 papers that were presented that week and over two different rooms, so I've listened to over 30 being presented and I felt a bit like a luddai in some ways, because I'd got a notepad and a pen and I was writing notes and I wrote notes on every single one. I took something away from each one that I listened to. But I chose AI text to image generation in art and design teacher education a creative tool or a hindrance to future creativity, and that paper was presented by colleagues from Oslo Metropolitan University. They set out with the aim of exploring how text to image generators can contribute and change art and design teacher education. They carried out some experiments using a range of image generators, including mid-journey Dali-E stable diffusion. They also collected data through observations and interviews from text to image communities and users.

Alison Hardy: 5:39

Can I just clarify there. So you've listed all the different AI generators and I'm going to put links into the show notes. That's why I went on mute very quickly and I'm catching myself up here. But what was interesting is that there's two things that I want to kind of. I know you haven't finished your synopsis yet, so this is really naughty of me interrupting. Already it's about art and design teaching. So I think we're kind of gonna be a little bit aware that that's a different context of D&T, but I'm sure we'll kind of explore that as we go on. But also, the participants were the presenters, weren't they? They did it to themselves. It wasn't with children, was it? It was. I don't wanna clarify that for people listening. This was teacher educators, academics, exploring, including having a play, really, weren't they many times? And that's not to dismiss what they were doing, but that's just that context.

Joanne Taylor: 6:32

Yeah, and I found that they will go on to talk about their conclusions or their findings and their questions Really, really interesting. I think this is one of the things that I've appreciated since the conference is that teachers are really busy people, and conducting active research like this is often something we don't have time for, because we're all capable, we could all do these experiments, but it's finding the time, and so taking 10, 15 minutes to read a paper where somebody else has had the time to do it is a good use of our CPD time.

Alison Hardy: 7:12

Yeah, yeah, absolutely, absolutely, and that's hopefully that's what people will get from then listening to this conversation, that they can read it, but hopefully we're gonna pick it and pick it a little bit more. So you got as far as explaining. They were kind of experimenting with these different AI generators of text to image, so can you talk a little bit about their process and just kind of give an overview of it?

Joanne Taylor: 7:35

Well, they did some, like you said, some experimentation themselves, and then observations and interviews with people that use it more as well. They also talked about the history of visualization and how traditional methods, such as communicating with drawing and painting, take time and skill. They suggested that the current trainee teachers will be teaching for the next 40 years. That's what we hope, isn't it? And as AI is developing at such a pace, it's impossible to predict what the future will look like for those teachers, and they're going to propose that teacher education must prepare students to facilitate future learning processes and challenges as a skill in its own right, rather than just understanding this particular development.

Alison Hardy: 8:29

Yeah, yeah, so there's some lessons to learn there. Isn't there about preparing, you know, the next generation of teacher educators as these new technologies come out? Because, as you say, what's gonna happen? I mean, it makes me think that artificial intelligence is a form of a disruptive technology, because it really is disrupting. You know as they talk about, don't they? It's disrupting whether we teach drawing, you know hand skills, or whether it's how to use these new technologies that are appearing. So there's kind of a number of different levels to it, isn't there? It's like it's a pedagogy, but it's also a design instrument, I would maybe call it, but also it's an emerging technology along the way. So there's kind of lots of facets to it that I think design and technology teachers might need to think about when they're maybe approaching AI. I don't know what you think.

Joanne Taylor: 9:34

Yeah, I agree with those points that you make. I personally, I think it's really exciting and I see it as being an additional tool as opposed to instead of. I mean, I guess, why is that? Well, it reminds me of a webinar I did with the Design and Technology Association, paul Russell from Loughborough University. He demonstrated some of the possibilities. I think I did it last year. He demonstrated some of the possibilities and explained how his undergraduates were using AI text to image generation and how they use it as a starting point, as inspiration, because it's a picture, it's nothing else. It doesn't give you any more information. It's a picture that's produced, and so the way that we, as teachers, ask our students to generate images is already not one way of doing it. So there's CAD, there's pencil, there's modelling. We already have a range of communication strategies that we employ and ask our students to do and develop their skills with. So I see AI as being complementary to that, rather than instead of.

Alison Hardy: 11:04

Right, okay, that's interesting. So do you see this? So do you see AI as a communication tool?

Joanne Taylor: 11:11

Well, ai is massive. If we just focus on the AI text to image which is what this paper was.

Alison Hardy: 11:18

Yeah, so let's keep the context. We're just talking about this.

Joanne Taylor: 11:23

Yes, I think it can be, because it's combining the prompts that we give it and trailing the internet for images and merging images together based on the prompts that we give it. So, in one way, I think it is a communication tool, because we're communicating the prompts from our heads, or from our students heads, into a format that we can all visualise as well. We can all see.

Alison Hardy: 11:56

Right, okay, so then it becomes that. So then it becomes. So you get things at the visuals, so then they can share their ideas with others and they talk about don't they in the paper about the importance of sharing, because they talk about the discussion boards, don't they, where people are sharing ideas about the language to use, as well as what the images?

Joanne Taylor: 12:15

Yeah, and that was actually the question I asked at the end of the paper. So for people who haven't been to part, the papers are presented and take about 10 minutes and then there's a 10-15 minute session of questions or comments or discussion and it's really you could argue that that was. The most valuable part of many of the papers presented was the discussion afterwards and the interaction with the people that have done the research. It was quite awe-inspiring really that many of the people presenting presenting a language that's not their own and they can also listen to and answer questions on the hoof in a language that's not their own. It's quite amazing. It's quite amazing yeah.

Joanne Taylor: 13:05

Yeah, really impressive, so yeah. So the question that's the question I asked is, you know, this comes down to language, that we use the prompts that we give AI and actually the picture or the collection of thoughts in a student's head or a designer's head perhaps isn't different, you know, depending on what they're, how they're trying to get it out of their head and share it with others, whether it's AI generated images or whether it's a drawing or a model, Maybe their thought process is the same and it's just a way of and therefore, I would suggest that it is a form of communication, a communication tool.

Alison Hardy: 13:51

Yeah, it's interesting that you talk about language, isn't it? Because they have some discussion, don't they around? And they put some screenshots, don't they, of some of the different servers and the chat communities where they're sharing and collaborating. And what struck me and I've used chat GPT more than anything else is the output is so much dependent on the quality of the prompts that you write, so you have to have language to be able to do it. So the strength of it is is the speed and the range of what it generates, which is the wow factor. I think that they talk about that. If you did this with children, they'd be the wow factor, but ultimately what it generates is restricted by the prompts that you put in. And if children don't have a good range of language on, you know different words they can use. That will restrict what's generated. Is that is?

Joanne Taylor: 15:00

that not the case that's my understanding, and he's actually an avenue that I'm considering for my research is the use of language and design and technology, or the use of language in creativity or something along those lines? Because it's, if we can't draw out of the students, the children in front of us, what they've got in their heads, then their ideas can't. We can't, you know, we can't see them, we can't share them, we can't help them to create whatever it is that they can think of. So, yeah, now I think it's a really interesting question the use of language in our subject and how we promote that and how we draw from students, how we draw things out of them and they don't have the actual words. You can quite often in my experience, get a student to show you, like with their hands, you know gesticulating what they mean by something.

Joanne Taylor: 16:05

And when modelling came in as a much more significant part of the assessment for GCSE in particular, it really led itself, lent itself to students who can model effectively and communicate their ideas. You know, students that can't draw, I imagine for us are similar to students in other subjects that have got the fear of the blank page with writing. If they don't have confidence in their drawing skills, then they think they don't have any ideas, but actually it's. It's our job. I see it as our job as teachers to draw it out of them in another way. So you know, giving them an isometric grid paper or giving them cardboard and scissors, whatever it takes to get them to share their idea and this is this text to image generation is just another tool I would suggest.

Alison Hardy: 17:00

Yeah, so I just kind of. I just want to go back to this thing about image. I'm curious Because to me, the children need to have, need to know certain words, have language, before they can completely they can make some use of AI, but it will receive, it will reach limitations, I think, because it is text and I mean, as you say, it may well be us pick up on at some point some gesticulation and other things and they talk about trying to find in the paper the wire factor. That's on page 705 and that the first attempts at creating AI images, they say, may be fascinating, but the wow factor will not necessarily last for long and trying to create something based on ideas and imagination may lead to disappointing results that do not match how one visualized the ideas in the first place. The natural limitations of a given technology or tool will limit the possibilities.

Alison Hardy: 18:01

But I sort of thought I missed something here. It's it's it's the knowledge of the prompts, which is what they explore in the discussion boards, I think, which is which is absolutely key to keeping that wow factor. I think, because the frustration will come is I've got it in my head, but as you're talking about, as you're gesticulating, and people won't see the video of what we're doing is we need to teach the children language to be able to help them articulate that, to be able to write the prompts, to then get the different, you know, unusual idea. Thinking what, what was the word that they use? I can't, I can't remember what they used as their you know, innovate, innovative ideas, I think, for example, and so, yes, I was kind of a bit surprised that that wasn't explored or pulled out of it might have been pulled out in the presentation, send it.

Alison Hardy: 18:58

People write papers, but then, when they present the kind of their thinking developed in this, yes, that was so. So you as a communication tool, and see, you've talked about, you've talked about the communication tool, the fact that children can, can use it, but it's about helping get the ideas out of their head. Were there any other key points from this that you felt were useful for the wider D&T community?

Joanne Taylor: 19:21

well, ai in general is a really hot topic and, as he's often the case in schools, we don't really know how to integrate it safely and there's the fear of the unknown and. But the reality is the students are likely to be more savvy with this than we are and and I personally see it as a huge opportunity rather than a threat. But we do have to go into it with our eyes open and, you know, embrace it but also be cautious. I would say, and I found the the analogy at the beginning of the paper between artists and the introduction of cameras and photography and and how photography was going to be the demise of art, and artists were quite upset about it and you know we sit here now we've got both art and artists and photography and photographers that they've evolved alongside one another. So I, going back to your point about potentially the demise of drawing skills, and I really do see that both AI and drawing skills can exist as complementary skills rather than on their own you touch there a little bit about integrating it safely.

Alison Hardy: 20:45

Do you want to expand on that point? Could you do talk about ethics a little bit?

Joanne Taylor: 20:50

yeah so. So my own teaching philosophy is are the kids safe and how have they learned anything? In that order, and not just from a D&T perspective, because we should always make sure that the children that we are in our care are safe. But, yeah, no, the the bit that I haven't read about or thought about before this paper and was the bias and the stereotype. So well, I've obviously thought about copyright and where. If AI is trailing the internet for information, then somebody's put it out there already. So copyright is going to be a question for everybody. But the bias and the stereotypes was something that I hadn't really thought about.

Joanne Taylor: 21:39

So if we're to introduce this into our classrooms, how do we do it fairly in a way that our students are aware that there is bias from what is found on the internet?

Joanne Taylor: 21:57

So the examples in the study are things like if ethnicity is not specified, then it's a white male. That is often the result, and if the input, if you ask for a female, then the response that you get is young and skinny, and that's something that we need to be aware of and we need to. I think our responsibility to our students is to make them question the information that they get back and to critique it and is it, you know, like if it was a source in history? Who wrote it? Where did it come from? Is it authentic? What's the bias? All of those things are part of the the skill of using a source in history. We need to. We need to do that as well. So if we're going to use AI in designing technology, our students need to be able to critique it and question it, and explain that to us and to the way they're assessed in their NEA or whatever.

Alison Hardy: 23:02

Yes, I think that's kind of where it comes back to.

Alison Hardy: 23:04

Something I was talking about earlier is about teaching AI as a disruptive technology, as a technological artifact in itself, and some of the unintended consequences of this new disruptive technology and I'm using the language is emerging technologies in the English National Curriculum, but Barlech, stig and Givens wrote a chapter in the Learn to Teach book about disruptive technologies and they talk about some of these things that are hidden within these emerging technologies that we're not aware of and that we need to bring them out to show children that there are these things to think about as technologies develop. And, yeah, I think that section that they write on page 706, where they talk about bias and stereotypes, and then previously on 705 where they talk about ethics, copyright and censorship I mean it's a short paper, but they do highlight that and I think it's not just is it for the teacher to know about that and then be teaching the children to think about that when they're writing the prompts, but then also maybe to be teaching it as part of them understanding this new technology.

Joanne Taylor: 24:16

Well, yeah, absolutely. I mean it's a little bit. It reminds me of when 3D printers came into schools and the students are very enthusiastic about it. And you know the same when laser cutters came in. Or the same when CAD became more prevalent, and you know, like the pro desktop years where everybody got really excited about having access to 3D CAD software, students are really enthusiastic about these types of things, but they come in with ideas that they're going to print something that they found, like on Thingiverse or somewhere, and for me that's not. I've always said no, you can't print something from Thingiverse. You can print something you've drawn. Let's have a go on, you know, on shape or whatever it is, to have it to draw something that's yours rather than print something that's not yours. But there's lots of open source materials, so that's an opportunity to teach about open source and about copyright and to bring that in as well.

Alison Hardy: 25:18

Yeah. So I suppose what we're kind of exploring here is that there's the excitement of this paper and it covers a huge amount in a very short paper, but there's lots for teachers to think about as they're thinking about their own sort of practice and what they might do with this. So there's a lot of video, barrows being sort of leading quite a lot of this, and actually she's been doing some really interesting work, and I'll put a link in the show notes to some of the stuff that she shared on LinkedIn.

Joanne Taylor: 25:44

Yeah, she's talking at a conference she's presenting next week. One of the things that I think is particularly interesting for others design and technology specialists is how the image generated doesn't give you any information about what it's made from, how it's made or even if it's possible to make it and our assessment criteria is still designing and making. So Paul Russell, who I mentioned earlier, from Loughborough. He explains how his students use it as inspiration and I thought about that.

Joanne Taylor: 26:33

I think, well, what my students use as inspiration, and it's things like it's called the work of others, but you know design classics or existing products. They often use those as inspiration and stepping stones towards whatever their product might become. So I could see a place where AI generated images could go in there, and then you know there's a huge number of marks for the students to gain, explaining that journey from this idea that's come about from wherever in this case AI to its reality. And how did they get from one to the other? The paper also goes on to talk about suggests that images generated by AI help students avoid design fixation, and I mean I certainly find that my GCSE students in particular are still plagued with design fixation. They can't think beyond one solution.

Alison Hardy: 27:46

Can you explain a little bit more about how they think AI can help counter design fixation, one of the things that they talked about was how?

Joanne Taylor: 27:59

because the ease of generating such an array of images. It takes away that need to have a high level or a breadth of skill and dexterity to produce it yourself. So drawing and modeling and sketching or CAD, they all involve a level of competence to be able to express your idea, whereas this is much quicker and doesn't involve the same level of skill because the computer's or AI is doing it for you. So all those thoughts that we touched on earlier when we were talking about language, all those thoughts that students have but they dismiss them before they get out of their head because they don't know how to draw them or they don't know how to communicate it. If they've got a word for it, they can put it into AI. So my understanding is that that means that we can create more and different ideas before we've put up the barriers to why we can't.

Alison Hardy: 29:10

Right. So it's about the speed of the generation, and they talk about it, don't they? The results that the AI generation gives means that it can happen quite quickly and rather than children thinking I've got to draw it out, I've got to label it, I've got to do this and that and the other. It's that kind of speed of generation, speed of visualization, let's put it, which helps with the fixation have I got?

Joanne Taylor: 29:45

that right? Have I captured that correctly? Yeah, I think so. And just the? It's not just the speed. I think it's the low stakes because this is just my own thoughts but the low stakes nature of it when we use the whiteboards for students to do their starter task. It's low stakes testing, so everyone can access it and you can put down your wildest and wackiest ideas, because it doesn't really matter if you're not quite right. This is this feels the same. It's not the fear of the blank page, because we can just do it again or we could just type in some more words and see what comes up next. It's the speed and the ease of access means that we can come up with so many different things so quickly that if we don't like what we see, we can just move on very, very easily. Yeah.

Alison Hardy: 30:41

Yeah, I think that's a really good way of putting it. Yeah, it's low stakes, but equally, it's kind of got to be contained, I suppose to an extent hasn't it. So you've kind of I mean, you've really kind of opened it up there. I think we've only sort of touched what AI text to text from a generation can do, but I think this paper is a really good starting point and thank you for picking it, joanne. So it'd be really useful as we finish Whether you can summarise, after a map we talked around, what a couple of points are that we're beneficial to teachers, researchers or other interested in D&T that they're going to take away. What are their takeaways from this, if they don't want to listen to the whole thing right, I mean you're talking and why they should do. What's their takeaways? Okay, so two takeaways. One regarding AI.

Joanne Taylor: 31:30

Be aware that there's a bias and stereotypes. And the second one is that research is not far away from educators. It's something that we can really benefit from. It can save us a lot of time because somebody else has done the experiments and the reading and the research and we just need to read what they've collated and concluded. Because teachers are so busy we can't always investigate and look into everything that we might want to and use that. But we can use it. So there's an increasing drive in education for research and evidence-based learning, personalised CPD I mean.

Joanne Taylor: 32:19

The last school I was in gave everybody a protected free so that they could do their own CPLDs, continued Professional Learning and Development. So I was lucky enough to be able to go to a four-day conference. It might not be possible to have a week off and fly to China for Pat 41, but it is possible to engage with the research community and, like I said earlier, it is a really vibrant community nationally and internationally. It's quite accessible to read a paper. The Pat 40 proceedings are all available publicly. It's quite easy to share your thoughts with what you've read with your team, with your local networks, school or trust-wide, but also to join in the conversations on social media carefully. Yeah, yeah, well, thanks very much for those two and thanks for sharing this paper.

Alison Hardy: 33:21

I'm picking this paper it's been a really interesting one to kind of get going and, as we've talked, it's kind of prompted some things I'm going to pick up in Thursday's episode that I'll be releasing, which is for subscribers, where I kind of explore some of the things that you've talked about and maybe link some of the bits of research and other writing for people who subscribe to be able to catch up with as well. But thanks very much for joining. You've got my brain going and I'm sure you've got plenty of teachers brain's going as well and hopefully come on the podcast and talk about your research at some point in the future. I'd be delighted. Thank you for having me.

Alison Hardy: 33:55

I'm Dr Alison Hardy and you've been listening to the Talking D&T podcast. If you enjoyed the podcast, then do subscribe, on whatever platform you use, and do consider leaving a review, as it does help others find the podcast. I do the podcast because I want to support the D&T community in developing their practice, so please do share the podcast with your D&T community. If you want to respond to something I've talked about or have an idea for a future episode, then either leave me a voice member via speakpipe or drop me an email. You can find details about me, the podcast and how to connect with me on my website, dralisonhardycom. Also, if you want to support the podcast financially, you can become a patron. Links to speakpipe patron and my website are in the show notes. Thanks for listening.