Back to articles


With AI being a hot topic, we turned to Kevin Walker to discuss AI and its implications for creatives. Walker, an Associate Professor of Immersive Media at Coventry University, leads a research theme, focused on AI and Algorithmic Cultures at the Centre for Postdigital Cultures. His research encompasses the intersection of natural and computational systems (more about that later). And with a background in anthropology, art, and technology, he explores his findings in theory and practice through software programming, exhibitions, events, and performances, as well as academic and creative writing.

Let's dive right in: do you think creativity is fixed and innate, or is it something that can be trained and explored? 

KW: "There are different theories that you could apply to this. There's one theory that is related to this idea of systems. Humans are a kind of system, and all of us are composed of all these other systems like the brain and the digestive system. But, at the same time, we're also part of more extensive systems, like societal and economic systems. Gordon Pask coined the conversation theory. The idea is that it takes two people, two actors, to interact with language and actions. But the key to that system is feedback. And so, the theory is that creativity always relies on some feedback mechanism, which could consist of thoughts we form in our head or feedback between two different entities. 

There's another theory by philosopher Daniel Dennett, who said that it’s not just people who create things, nature does as well. He looks at this in terms of design, so humans are intelligent designers, but according to Dennett, nature also undertakes this creative process through evolution. And that kind of creativity equals competence without comprehension. 

So where we, as humans, might be able to reflect on and comprehend the things we make, nature is doing the same thing through trial and error. Without, as far as we know, any intelligence or comprehension. In that definition, nature is also creative, and some people believe that all our creativity is just mimicking what happens in nature."

Does that mean that animals are also creative? 

KW: "In that definition, yes, because they're adapting to their surroundings and to things that happen to them, either during the lifetime of one animal or over a more extended evolutionary period. Another philosopher, Emanuele Coccia, thinks of this in terms of plants. Each plant is not only adapted to its specific environment but can also change that environment in some ways by creating flowers, releasing pollen into the air, etcetera. And the same applies to humans. We're not just part of our environment; we can also change this environment. We can't, however, be separated from this environment. Think of fish. You can't take fish out of the water; it's their environment. The same goes for humans and our need for air. When we breathe in, we're taking in some part of our environment, of the world around us. When we breathe out, we give back some part of ourselves to the world. So in that sense, we're inseparable from this world. 

Coccia argues that taken from his perspective, a fish is not so different from the environment it lives in. A fish is merely a different version of the water, a more solid version. It becomes interesting when you think about humans and creativity and this idea that intelligence is not just something inside our heads but distributed in our environment. For example, when I have to remember to take out the trash, I place the trash by the door. That way, I don't have to think about it anymore. The information is now in the environment. It's not in my head anymore. So that raises the question: how can we be creative by responding to and changing our environment? 

And to return to this idea of systems: look at a company like Google. When companies like that grow to a specific size, they cannot be so creative anymore. And so when they need new ideas, they have to hire people with ideas or buy smaller companies doing interesting creative things. They're bringing in creativity from the outside and introducing it into their own system."

You're a creative person using different means of expression, from theoretical writing to poetry. How does that work? 

KW: "I don't always intend to write poetry. Sometimes it just comes naturally. The same happens with ideas for an article. The best time for me to write is in the morning when I'm fresh. A Ph.D. student of mine said the best time for him is in the evening: after a couple of glasses of wine, he can be more creative. We're taught to work according to the clock, and in reality, we all have different stages of the day when we're more productive or creative. And those stages can differ per person.” 

Left: Leonid Pasternak, The Passion of Creation. Right: Modern version of Pasternak's work made with AI.

The technological world of AI and the world of poetry seem so far removed, but in your work, they aren't, are they? 

KW: "So I'm working on this current project where we use AI to help us structure a narrative. And we tried to train the AI system using poetry. So I copied and pasted these poems into a text file. A lot of poems. And I did this for a whole day, not even reading the poems. And I was looking at these poems all day, and by the end of that day, I felt like I was seeing the world as a kind of poem. Everything I saw, every object and every action: somehow, it came to me as a kind of poem, which was interesting. As if I became programmed by poetry. It’s not unlike what writers say: if you want to be a writer, you have to read a lot.”

It’s about input then. Which, brings us to AI. How do you think AI will effect creativity? 

KW: "I currently only use it to generate some unexpected connections, some randomness or some noise. Let’s go back to Alan Turing's original paper on Computing Machinery and Intelligence. In it, he talks about a classroom. Students can learn what a teacher wants them to, but true intelligence, and maybe we can say the same for creativity, is when we introduce something new, something different, something random. He says that for a machine to create something new, you must introduce some randomness or noise into the system. 

Some say that AI can never create anything new because it only knows what has been put into its system. But potentially we could train AI systems, which is what I’m currently looking into. Can it create something new and interesting that we wouldn't come up with ourselves? I urge my students not to just put in a prompt and then use the images generated, but instead use those images as a starting point. ChatGPT, for instance, is not very good at writing, so you can use it, but only as a first draft. You will need to do something else and use your knowledge and intuition to shape that text into something else. So I think it should be a collaboration."

Is human creativity still relevant in the future? 

KW: “If you define creativity or intelligence in a very narrow way, it's just a mental process that we can model somehow using an AI system. But to me, creativity is not just inside the head. It's not only a mental thing, it's also embodied. And that's where the difference is. So AI is good at doing digital things, things that exist on the computer and the Internet. But we live outside of the digital world as well. 

When I'm drawing something happens to me that's not just mental. It's the actions that I'm undertaking with my body. So we will always have this difference from machines in that sense. Now if machines were able to execute code and have a kind of memory, then it can start to have some actions in the world. And if it can execute code, if it can iterate and improve on its code, if it can start to be autonomous somehow, it can begin to impact things happening in the world.”


“To me, creativity is not just inside the head. It's not only a mental thing, it's also embodied.” 

Some people say the main difference between AI and humans is that humans have the urge to create an AI. But now you’re telling us that AI could have some autonomy. 

KW: "Yeah, well, the question is: can it have its urges? Can it have desires? I don’t see that happening. Then again, I spoke to a robotics engineer some years ago who created this super simple and tiny insect robot. It only has like two or three components. It has a light detector like a photocell and a capacitor. To store up the charge that it collects from the light. And that's it. And then the little motor moves it, so all it does is where it detects more light, it moves towards the light. So so this. Super simple system. It already has some autonomy because it's seeking out the energy, the morning, so it can perpetuate itself to survive. So maybe there's something there? 

There's one theory of creativity by this guy named Graham Wallas from the 1930s. And he said that you must first collect all of this stuff. All these visual references and things. But then. What you do is go to sleep. And you know if you're working on a problem late at night or something you need to sleep on it. I do this with programming. I'll be trying to figure out what the hell is wrong with my program. And then I wake up in the morning and realize I was missing a semicolon somewhere. 

Now I asked ChatGPT about this. I said sleep is essential for creativity, ChatGPT agreed. I asked for some research to support this claim. ChatGPT said there was a 2013 paper about this and another paper, but it didn't give the exact references. Then I asked if it could become more creative. Could you sleep? Do you sleep? And it said, of course: ‘I don't sleep. I'm an AI system. I'm just running all the time.’ And I said, well if you could sleep, you could be more creative. But I don’t think he got it, haha."


    Sign in to reply and see more comment(s)