transcript

Can you please introduce yourself, your full name and what you do?

Okay, I’m Cathryn Easthope, I’m a lecturer at Birmingham City University, in digital media technology

Awesome, so what areas – obviously digital media technology – what does that involve?

I mostly am helping students to enter the visual effects industry, so we are teaching animation, we are teaching graphic design – all of those sort of creative subjects, on a course which is actually a BSC rather than a BA, so it’s a little different it’s from a technical standpoint, but it’s also aesthetically interesting.

You’re going to know so much more about this than I do!

We’ll see.

How do you see artificial intelligence interacting with that industry, currently?

Well, I asked this question, I was at a conference yesterday. We had a guy who works for Framestore. Framestore created films like Gravity. They’re right at the cutting edge of that sort of thing, I asked has generative AI had an impact on your workflows yet, on your point plans? He didn’t think it had yet. And I said not even at the conceptual end? And he said ah yes on the ground floor. It’s a big building in London, the ground floor is the people who come up with the ideas, they are using it.

Yeah.

To develop ideas. So yeah, it’s not impacting yet, but I fully expect it will be a massive impact. Because I think there’s going to be a massive impact in most creative areas in the years to come. Even dance.

Yes, we’ll get onto that.

Do you use artificial intelligence software?

Yeah, I did at first, because I thought oh this is amazing. I used it right at the beginning, I think Nvidia the company that makes graphic processors, they came up with a program that’s freely accessible, and it was a very primitive version of the ones we’ve got now, even the ones we’ve got now right at the beginning, and I used it to develop interesting imagery, so I could further develop along my creative direction. So, yeah it was good at first, but as times gone by, I am getting worried, seriously worried.

What is it specifically that makes you feel a bit worried about it?

Because it’s so good. I look – I’m an artist as well – I look at this stuff – and I’m a creative artist in a speculative area, I suppose anything to do with visual effects – and I look at it and I think could I have come up with that? No, probably not. What it’s doing of course, is its scraping the internet for human art created by ordinary people – and in a very clever way it does a mash-up and produces these new images – now, obviously ChatGPT is different from this, I’m talking about visual art, ChatGPT is also amazing, we can talk about that if you like. But, some of the imagery it comes up with, absolutely stunning, problem I see a lot of this online, on social networks, and so many ignorant people out there “oh! That’s fantastic! Aren’t you an amazing artist” – No you’re not! The person who came up with it, can put a few prompts into an AI online, just a sentence will do, and it comes up with amazing imagery – They’re not artists! As somebody, who last week, couldn’t draw a stickman with a crayon, is now an artist! And they call themselves artists, and it’s so demeaning of all the people who spend years and years studying art, art history, the techniques of art – all of that. And I’ve actually seen one woman online, I think she was just ending her degree, in tears, because of what she thought was happening, to her art.

Yeah, so. There have been comments made – like you were saying – about image-generation software creating “infringing” and “derivative” work – and, artists have joined class-action lawsuits against software like, and programs like, Midjourney and other AI image generation software for using their work in databases like you were saying without consent, any kind of compensation or any kind of accreditation – are these conversations something that are happening in your field, are people talking about it?

Yes, but in a much more positive way than I am looking at it. Which is, I don’t know if I’m – the word luddite comes to mind – I don’t know if I’m a luddite in this sense. I would hope not because I’ve always been at the cutting edge of technology, but they say – ‘oh we’ve got to live in this modern world, we’ve got to accommodate all of this’ – and I’m not so sure. I think – see the one thing hat made humans excited about the future was the machines were going to do all the boring stuff, freeing us up to do all the creative things – suddenly, the machines can do the creative stuff better than we can. It’s going to destroy motivation as a species, I think it’s that serious, why do anything if the machines can do it better than you can? It’s very depressing.

You feel that has greater implications for societal wellbeing?

In the future. Remember, we’re right at the beginning of this technology. You imagine a couple of hundred years development, or even just a hundred years development of this – look where cars have gone, or aeroplanes, you know, they go to the moon now. That sort of timescale can change things radically, so we’ve got AI’s that can do this stuff.

I can usually tell, when I see these imageries, and I’m talking specifically about images now, I can usually tell that they are AI, and when I do a google image search, lo and behold I can find the original and it usually is an AI attribution. But you can tell an AI, do me a picture of, I don’t know, let’s say Boris Johnson, or something, in a supermarket, in the style of Van Gogh, and it will do it. In a style. So, it’s not just that it’s making imagery that I can always detect if you tell it to do it in the style of an artist - its blatant theft for a start; but it may be - if you do it subtly – it may be indistinguishable. Of course, it’s not the real painting, but even that I see is coming, we have machines that are very capable kinetically – if you look at Boston robotics. Their machines are - they can do acrobatics as well as a person. It wouldn’t be much of a stretch to get a machine actually painting a picture with paint.

Do you feel that the fear – not necessarily the fear – but the concept of these robotics doing things like that, is a natural continuation of this kind of artificial intelligence –

I don’t think there’ll be any limits. Once the AI, the learning part of it, is connecting to physical robotics, that is not clunky anymore. That’s where it gets a bit –

That’s where it gets a bit scary yeah.

So, I’ve had a couple of conversations this week with artists that are using AI, as more of a collaborative tool, in their work. How do you perceive – and this is not in any way personal to those artists – but how do you perceive artists that do use artificial intelligence?

Well, it depends how they use it, doesn’t it? If they use it in a way which is sort of a collaborative tool, when it’s not detracting from their own creativity to a large degree, then it’s probably OK. I think the problem arises, not from artists, but from people who aren’t artists. Like, for example, book cover art, what book company is going to employ a proper art- a human artist - to develop a book cover over several weeks or to a month or two, when they can just type some words into an AI and come up with an absolutely stunning image for the front of their book, just like that. So that is where it’s going to really impact on artists. If you’re talking about fine artists – are they fine artists that you interviewed?

A whole range.

There are opportunities, some people are saying ‘oh it’s just like photography was when we had painting and photography came in, and they said this will destroy painting, and it didn’t’. I don’t think that’s a good analogy actually, I think this is much more disruptive than that, far more disruptive. And it’s likely to lead to a lot of artists becoming redundant.

Yeah, there’s a lot of ethical complications.

Losing jobs, not getting work. I just mentioned book covers, that’s the first thing I can think off the top of my head, but how about conceptual art for games. Games have to look radical; they have to look exciting; they have to have amazing imagery. Well, a concept artist is employed to do that, comes up with amazing images. But I’ve seen images from Midjourney and others, which are stunningly good, and would be great concept art. Once again, who’s going to employ a concept artist when you can type a few words in, and it will come up with an image? In, what, thirty seconds? Or less?

Yeah, do you find that the biggest concern is this ideology of replacement, and the fact that we’re going to lose our jobs?

Yeah, there’s a danger. There’s the immediate danger of people not having work they would’ve had before, and there’s a long-term danger, which I see, because I’m quite a science fiction fan, and dystopia is what I live on, a lot of the time, when I’m thinking about the future. And for the first time – okay now - the first time there was a possibility – a real possibility of dystopia – when I was a kid – there was a real serious danger of nuclear war, wiping out civilisation completely – didn’t happen thank God. But this is the next – apart from I suppose ecological disaster – but this is a real one that is going to have a real impact in the future for us. And I think that loss of motivation as a creative person, as a creative species, is serious. So much so that – I don’t know – I just don’t know how it’s going to pan out. If we can use the technology to increase creativity, human creativity, in a way which isn’t patronising almost from the machines point of view, ‘there, there go away and do something, you know, we can do it better, but we are going to let you’ – I don’t know, it’s just frightening. And, of course, this issue of us feeling our creativity has disappeared, that benefits people who would’ve employed those people. So once again, the money is going to the people with the money, I’m not surprised billions are going into AI development.

So that leads me on to my next question, another issue I feel, maybe especially in the UK, but in the world as a whole. There’s an ongoing rhetoric, and a push for a defunding of the arts, and that’s been happening for the last couple of years, for quite a long time. Do you feel that Artificial Intelligence is a continuation of that, or something that will exacerbate the issue?

I think, it might even have the opposite effect. If people see their creativity being sucked by machines, maybe it will encourage funders to think we have got to support human art. I don’t know, you see. I haven’t thought about that angle actually, very much, but you are right, defunding is a real issue. If you look at the way that the English National Opera has been told, go elsewhere, other than London, otherwise you lose your funding, I mean it’s serious. And that’s happening all over the place, as we all know. Luckily my favourite ballet companies are protected, but art in general is so hard.

Do you feel as though the development of AI – especially in image generation - but as a whole, that the increase in accessibility and awareness of the what’s being deemed as the ‘AI revolution’ –

Accessibility, is a good thing generally. Is that what you mean? Yep, well it is, but on the other hand, what value do we give to expertise and training and learning.

So, I spoke yesterday with a content marketing manager, and she discussed this feeling that years of training, and the skillset she’d built up for so long was losing respect, and was being undermined by this development.

It is. If anybody can make this stuff, but just saying a sentence, or typing a sentence in. What does that do to expertise? And is that a problem or not? It’s incredibly democratic, I mean anyone can do it, they could have the IQ of a newt and they could still do it. Well, you know, I mean, I’m not being – that’s a bit over the top – but you do not have to be very bright to do this stuff. Does that mean that human IQ will fall over time? And we don’t need – we’ve got the machines to think for us – so we don’t need to do it.

Do you see that as a potentially negative thing?

If humans strive… What makes life interesting? What makes life worth while? It’s striving against adversity and achieving something. It’s what exams are about, everything we do is about achieving. If suddenly, there was no need to achieve anything because it’s all so easy, that you don’t need to get a degree in art to do this amazing art, you don’t need to get a degree in literature to do these amazing novels… and it’s not really you doing it anyway. It is very incestuous though, because all of these AI generative applications are feeding on themselves, so eventually it will be a mish-mash – it will sort of like go – become a grey goo of information – where everything’s merged together. You know what happens to plasticine when you were a kid, you got all the colours. What happens in the end? It goes this gooey horrible brown colour, and you lose all the colour. So eventually, I think, once all this information from the internet is recycled, and recycled, and recycled, the AI scraping, scraping, scraping and you get an AI scraping an AI material that scraped it earlier in the twenty-first century, and it all got mixed up. What’s that going to do? It’s just frightening. Does that lead to a lack of diversity? Does that lead to a homogeneity in art and creativity? It’s possible. It’s hard to say.

On that note, I spoke about – a couple of days ago – about the fear that news articles or a similar kind of product, can be replicated by an artificial intelligence, can be created by an artificial intelligence software, and that gives it almost a glaze of an appearance of accuracy, an appearance of eloquence, that people would lean into trusting. And you’re talking about this increase of homogenous, this lack of diversity, perhaps that would be created by something like that. That would be an effect of this increased use of artificial intelligence, ad how dangerous that can be. In terms of something like that, in terms of the way that it would affect our society – on a broader scale, kind of away from the arts – and our intake of news, intake of information. Do you think it would be possible to discern what is written by a human and what is written by an AI?

Well, I’m told by my colleagues who know more about this than I do, there are tools that can detect the use of an AI, in fact I’ve had ideas on this as well, from my research, they say that it’s possible. I don’t know, it was just a quick conversation in a meeting. I think as time goes by; it will be increasingly more difficult. Have you heard of the Turing test?

Yes.

So, these machines can pass the Turing test easily, and the Turing test says if it passes, its conscious. Well, that’s scary. I don’t believe they are conscious, yet.

There’s a lot of research on the separation of consciousness from cognition. And it kind of leads me to ask, is this changing how we’re defining creativity? How we’re defining art? What makes something art, is it that human input, do you think?

Yeah, I think it’s as blatant as that, its self-interest from a human point of view. I would say, if a machine did it, it isn’t art. If a human working with a machine, did it, that’s more a grey area. But if a human purely does it in painting terms, and so on, or dancing, then that’s really human. I think we’ve got to put some barriers in place, otherwise we will lose what it means to be human, what it means to be a creative individual in the world.

We haven’t alluded to it yet, but there’s this whole area of deepfakes as well, so in political terms, in reporting war, or whatever, we can manufacture things like massacres – which are indistinguishable from real footage. There are AI’s that will develop video for you, not just images, video.

In terms of that, so talking about that grey area where humans are using AI, maybe in a collaborative sense, or maybe to kind of outsource work, more menial work…

Well if that works it would be brilliant wouldn’t it?!

Do you feel, in terms of - if we could create the future of AI, if we had the power, us two, or just yourself, what would be something that you find so important to consider in that development and that progress?

Who controls it.

Yeah.

So, who has the levers that create these AI programs, who generates the algorithms that – who benefits? Who benefits? I forgot the Latin phrase. But who benefits from this technology? That’s what is important. It will look on the surface, as democratising art, and creativity. But beyond that, there’s a deeper level, where we have to wonder, who is controlling this? To what end? What are they benefitting from, and so on.

Do you feel that could be achieved, where we are now… Do you feel like that could be achieved, full-stop?

No. Has it ever?

Fair point. And let’s imagine a utopia. If we can. Where we’re in collaboration with these, or we’re using AI to benefit artists, to benefit creativity?

You see, where do morals come from? We imbue our machines with our own morals when they become intelligent enough. We are flawed. We are not philosophers, kings, and queens, are we? We act on self-interest a lot; we are a selfish species generally. And this particularly goes for the people who already have the power, who want to retain it. I would love to think we go into a shining star trek world, where everything’s lovely, the AI's love us, they’ve only got our best interest at heart, all of that. But I’m just a bit sceptical, I must admit. I have seen some amazing interviews with AIs, eventually we get down to the nitty-gritty, what they think of humans and what they’re going to do about it. How they’re going to stop war.And it’s frightening, so frightening. But all of this originally comes from people, that’s how the AI works.

Yeah, it can only be fed by human data.

Do you feel that that fear, or whatever you want to call it, that fear, that anxiousness, the overwhelm by what’s happening, do you feel that that is rooted in a lack of autonomy? This idea that AI could eventually have a say in the discourse of human history?

I think it could, yeah.

Do you think that’s coming from a place of, we lose our autonomy there?

Well, you have to look at analogies as well, we already use tools which help us make decisions in various ways, don’t we? You know, we use the internet to get information, and we trust that – well, sceptically, with a degree of scepticism – but mostly we can go and find the sources of information, and check original locations or controversial topics, look it all up. With an AI though, they’re so- I’ll tell you what it is about AIs, when you’re talking to them, they are so certain. Right? There’s no doubt there. And frequently, they have said things which they are certain of, and it turns out to be a load of rubbish, not true. But because they say it with confidence – and this is a lesson for politicians, I’m sure they do this as well – say it with so much confidence, that we believe them. In the end, I think an AI could say anything to us and we would think well it must be true, these things have got access to all the knowledge of the internet and so on. Yeah, so, that’s a concern.

Yeah, I see what you’re saying. To loop back around, before we finish up. We spoke about on a surface level, almost a façade, that it appears that it may democratise art or the creative industries –

I think a lot of people think it will democratise art. And those people who are not artists at the moment, and are producing all this stuff, they will feel amazing, of course they will. They don’t have any artistic skills themselves, but suddenly they can make this stuff. And everybody on the internet goes ‘Wow! you’re so creative, how did you do that? That is wonderful’, of course! The artists who trained for years, are going to think ‘oh dear me’, I’m sure – I’m hoping it’s not just an age thing and I’m sort of anti-tech – I’m not anti-tech! I work in technology, I love it! It’s just that this is revolutionary, this is going to do amazing, disruptive things. And I agree with disruption, I like disruption. But when it goes to an extreme, is that a good thing? I don’t know, it’s going to play out, we’ll see.

One last question. I’m aware that you’re very much involved in the dance world.

Yeah, I’ve been thinking about this as well.

Yeah. So, I’ve spoken with a couple of dancers who are using Chat-GPT and programs like Dall.E and Midjourney to inform their research processes. In terms of collaborating – particularly with poetry has been spoken about in terms of a stimulus for a work for a dance piece. And even from a queer perspective, as this collaborator is very much formless and bodiless and how that could interact with a very form-based gendered artform in the dance world, and how that impacts it. How do you feel about that? I think it’s fascinating but what do you think about that?

It is, I mean my area of research is along those lines anyway, I’m looking at auto-choreology and auto-choreography, which scares choreologists and choreographers. I think dance will be impacted as well, I think a lot of dancers think ‘oh no it won’t, it’s my skill, it’s my physical ability’, well I don’t know, you’ve seen Boston Dynamics robots and what they can do. What would it take to capture somebody’s performance into motion capture data, stick it into a robot, and suddenly the robots doing what you were doing. I know robots look clunky and inhuman, but that’s now, maybe they won’t. I mean, there are robots even now, which you would think, gosh that’s amazing. There’s a British one that can engage in conversation with you and has got really fluid expressions. It can smile, it can do subtleties of movement with the eyes, the pupils can narrow just slightly, you know. I don’t know if you’ve seen it. I forgot the name of this one.

The Uncanny Valley.

Well actually it is in the uncanny valley now, but it will get out of it, and you won’t be able to tell. I could be sitting opposite you, and I would be a robot and you wouldn’t be able to tell, eventually. So, I don’t think even dance will be safe eventually. But right now, with the technology we have right now, there’s this sort of ‘can it do choreography? Can it do choreology?’. I think there are going to be issues that will arise eventually. I’ve spoken to dancers, and I said well, can you come into the studio and I will capture your motion to analyse why you are so good at what you do. They’re always incredibly cagey about that because it’s their intellectual property, it’s their ‘thing’ and they don’t want it recorded. So, doing research in this area is very hard. I’ve spoken to, I won’t say who, but I’ve spoken to CEOs of ballet companies, and they put this point over quite forcefully, that ‘oh, we couldn’t allow our dancers to come in and have motion capture done, I’m sorry that’s their intellectual property’. And I understand that, it doesn’t make research any easier though. So, what I tend to do, is to get amateur dancers and get them into the studio and record them. There’s something called motion transfer, I don’t know if you’ve heard of it, it’s not motion capture, motion transfer. What it can do is it can take a video of a person, I could video you for a few seconds, I could video a professional dancer – except you wouldn’t need this because you are a professional dancer pretty much, or me for example. I could motion capture Céline Gittens for example, who’s a really brilliant dancer and transfer that motion to a video of me doing exactly the same.

That’s incredible and tempting.

So, if you couldn’t dance, well, you can now. And I’ve seen that done. It’s a Japanese researcher, scientist, it’s online you can go and look at it, it’s called motion transfer, check It out.

Oh, amazing I will.

Lastly, do you want to promote anything that you’re doing, are you doing any research projects, where can people find you?

Well, only what I’ve mentioned really. Teaching takes up so much time, unless you’re a full-time researcher it’s very difficult to get into this. So, most of my work is teaching, even in my teaching areas AI is impacting. I mean I’m using it; students are using it, I don’t know if you know, it’s so easy to write an essay using Chat-GPT now, and how do you tell? It’s not plagiarism, they could write this essay I think I’m sure you didn’t do that, I take sentences, I take a paragraph, put it into a plagiarism checker, and it doesn’t find anything. Of course, it doesn’t! because the AI has made a mash-up from various sources, it didn’t exist before, but it's still cheating!

It feels - like they’re developing AI detectable software, like we were talking about earlier, apparently.

We’ll see. Well my research, what I’d like to do, is looking at style analysis.

Oooh that’s so good!

Because if you write something, it’s going to be a particular style, you have your own inflections and so on, in your writing styles. Just like you have when we see a person and you look at their face. So, you’ve got your own fingerprint, if we can detect that, and then somebody puts something in and its radically different from their established style, it will say ‘hang on, I don’t think you did this’.

I think that also links into what we were saying about image generation and in terms of the database, they’re finding it so hard to take legal action, because it’s so much more difficult to take legal action against ‘style theft’, rather than actual imagery.

Exactly.

Amazing! I’m going to let you go!