AI and Art: The Limitations of Algorithmic Creativity with Nouf Aljowaysir

Kate Bevan: Hello and welcome to this episode of the Infosys Knowledge Institute's podcast on being AI first, the AI Interrogator. I'm Kate Bevan of the Knowledge Institute. And my guest today is Nouf Aljowaysir, who in her day job is a product designer at Medium, the blogging platform. I think she's actually one of the most interesting people I've ever had on the podcast, so I'm looking forward to talking to you, Nouf. Despite her tech chops as a computer scientist, she's also an artist and she's an artist using AI. That's not using AI in the sense that worries creators of using AI to replace human creativity. But she's using it to understand how intelligent tools interpret the world. We'll talk a little bit more about that as we go, but one of the most interesting ways to see what Nouf means by this is to look at her video, which we will give you the link to in the copy introducing this.

Nouf, thank you very much for joining me today. It's brilliant.

Nouf Aljowaysir: Hi. Thank you so much for having me. I'm excited to be here.

Kate Bevan: Can you tell me a little bit about your background? Because when we spoke, you talked about family history in the Middle East and your time learning to be a computer scientist, and it's such an interesting background.

Nouf Aljowaysir: Yeah, it's all over the place, but basically, I was born in Saudi Arabia, and I lived there until I was 13 years old, and then my family and I decided for me to study in the United States. So, I moved here alone and lived in a boarding school since that time while my family still stayed at home. So, I've been kind of living between two cultures since that time. And I went to quite a technical school called Carnegie Mellon University, which is maybe number one or number two top universities in the US for computer science. And so coding and computer programming was almost in every major at the school. So even though I didn't start with computer science, I eventually got there, and I just grew more and more interested in it. And yeah, that's kind of my journey.

Kate Bevan: When did you start working with AI?

Nouf Aljowaysir: After Carnegie Mellon I went to a master's program in New York City called ITP, which is, the best way to describe it, it's an art school for coders. It's basically, I expanded my toolkit, I learned more programming languages, and I basically used code in a non-utilitarian way. It was just to make stuff that didn't make sense, to make stuff that was funny or made you think, or made you criticize the technology. So, from there, it expanded my horizons. And then the first job that I had outside of ITP, outside of graduating, it was at an advertising agency where my boss just really wanted to work with AI. This was around 2018, 2019.

Kate Bevan: That's actually quite ahead of the curve there as well.

Nouf Aljowaysir: Yes, people would look at me and say, what is AI? What are you working with? So, it's crazy to see now people understanding what AI is, what LLMs are. Algorithm is just part of our natural speak now. But back then, 2018, 2019, no way. People thought I was a wizard of some sort. And so, he, my boss, really wanted to work with AI and he wanted to use it to create images to rebrand a client, to rebrand their visual identity. So, you can probably see it up now, the client is called ADP. I looked at a lot of generative techniques that were really popping up at the time, and I used those to generate all the visual assets for that brand. And from there I just got obsessed and I got more interested in it.

Kate Bevan: So how did you end up with a residency at Somerset House here in the UK? It's one of the very prestigious art institutions in London.

Nouf Aljowaysir: So, from that role, I wanted to work with AI in an ITP art school kind of way. And I was interested in identity given this background that I told you about. Well, first I did a residency with ThoughtWorks Arts in New York City, and they were looking for artists that were studying synthetic media or media generated by these algorithms. And so, I submitted my proposal that I wanted to look at identity, and I didn't really know where I was going, but it was a very process driven residency. So, I kind of was able to know more about what I wanted to create or how I wanted to really understand this huge umbrella of identity of people asking me where I'm from but feeling like it's very hard to answer. Yes, I'm from Saudi. Yes, I've lived here. But where do I really fit in? And feeling like that answer didn't make sense. So, I wanted to study it from a personal perspective, but also from this AI perspective. And so, my project really started there, and at Somerset House I wanted to create a film around that process.

Kate Bevan: I think just to describe your film a little bit for listeners. What it does is it takes footage, your family footage, and photographs, and it puts recognition technology on top of it. So, it's identifying your traditional dress as like a robe or a dress, and it's just getting it very, very wrong actually. I think one thing, the way you are using AI to create art is very different from what worries people about how we use AI to create art. I mean, do you think it counts as art if you use AI to create it?

Nouf Aljowaysir: I think yes. Why not? Because art is such a large term. What is art at the end of the day? It's a combination of influences. It's a combination of things. So, I don't see what's the issue of that? Is it good art, that's another question.

Kate Bevan: Well, how do decide what's good art, particularly when AI has been used to create it?

Nouf Aljowaysir: Well, yes, that's when I start to feel a bit sketched out about it. I understand that, yes, as humans we're inspired by other people, but often we use critical thinking skills to reframe it the way that we see the matter or how our inspiration takes place. Whereas AI is more like a copy-paste situation, which is why it can be problematic.

Kate Bevan: I mean, I watched your video all the way through earlier today, and I think it's actually really moving to watch because it feels very personal to you, but it also has really big things to say about AI, what we're using it for and where it's coming from. I think one of the things that really struck me, and I wanted to ask you about, are you worried that AI is coming out of the West? It's a very western paradigm in terms of the language that uses visual and written language.

Nouf Aljowaysir: Yes, for sure. That's what my film tackles. Just my whole identity in the film is completely reduced or generalized, or there's western labels inserted on top. So, it's clear to me that if I'm trying to look for who I am, and as kids these days are using this technology more, conversating with this technology more, it's important to ask, am I in the data set? Am I being represented or am I kind of reduced to this western view of how this culture is viewed?

Kate Bevan: Other podcast episodes I've talked with other brilliant women about how we tackle bias in datasets. But when you're using it the way you have to really draw that out, what's your take on how we tackle that?

Nouf Aljowaysir: I think that breaking down the monopoly, like Silicon Valley right now has a huge monopoly over these AI algorithms. And OpenAI, they're creating tools that are globalized and just used everywhere. I think that's really incredibly problematic because there's going to be biases embedded with that no matter what. So, I think if we decentralize a bit of that control and put it in the hands of communities or maybe artists, not necessarily these decision makers that are sometimes very tied to politics and governments. So maybe in the hands of creatives and artists, we can create a bit more of a fair system that represents that community a bit better.

Kate Bevan: For me, it feels like AI flattens everything down to a very homogeneous. And if you look at some of the image’s AI creates, and I'm not necessarily going to call art, well, we can talk about that later, it feels like it has a very flat genus look. And what I was really struck by in your video is how you bring out your own background, your own cultural influences.

Nouf Aljowaysir: Yeah, that's kind of what I juxtapose in the film. It's like the flatness and the AI is depicted as quite stupid because it is in the film against oral storytelling, which is a lot of depth and poetry. And in the film, what really, I juxtapose also is the poetic way in which we describe ourselves as human beings of who we are and how we've lived versus this categorical simplification of AI of you are this, you are that. And it's kind of interesting how intelligence is being defined in this way and not in the way the stories have traditionally been told.

Kate Bevan: When you were making the film, did you get it right first time? Was it delivering what you expected and how that built your film?

Nouf Aljowaysir: It was really hard to make this film because I think it was quite... I knew I wanted to speak to an AI, but there was no precedent for me to look at or use as inspiration. I kind of felt like I was inventing this... And in the beginning, I didn't even call it a film. I said, this is more of a visual diary. I don't know if there was a right or wrong. It was really about listening to intuition and listening to how am I depicting this juxtaposition in the best way possible? So, it was just trying different things to make sure I'm really nailing down that kind of understanding that... Yeah.

Kate Bevan: And do you feel that the end result is what you wanted it to be or what you expected it to be?

Nouf Aljowaysir: Definitely. The film itself, honestly, it was quite therapeutic to me because these answers I've grappled with personally my whole life. So, it was a therapeutic process to ask... Because at the end of the film I really answer where am I from? I do answer, how do I answer this question, where am I from? And I felt like the end I had to really personally introspectively look at how I would answer that question. And so, it was quite a vulnerable process and I'm kind of happy about the output because I felt like I reached an answer for myself even in my own life.

Kate Bevan: Okay, I'm going to come now to the way creatives worry about AI. I mean, do you feel threatened, by the way AI appropriates simulates human creativity? And also do you understand why other creatives also feel worried about it.

Nouf Aljowaysir: Yes, definitely. I mean, people are using it in ways to cut jobs, to maximize profit as much as possible.

Kate Bevan: Thinking about how you used AI back in 2018 in your advertising job, I mean arguably you were using AI for precisely that reason, to shortcut the creative process, or were you not?

Nouf Aljowaysir: Yes, in a way... That wasn't the main objective. The main objective at the time was to work with this really innovative technology to, I don't know, produce something that felt visual, that felt a bit unique and different. So, the objective at the time was that.

AI could not have done this without me. And this is what I remind people a lot at the end of the day, is that AI needs creatives. It does not have an imagination. It does not have a sense of creativity. So, it really is a back and forth and it needs to work with creatives, and it's absolutely nothing without us. And so, it can be worrying about certain applications, but without that training, it's really nothing.

Kate Bevan: That's really interesting actually because so often you hear a creator saying, it's going to do me out of a job, and I say when I'm talking to people, but you can't have AI without humans, and actually some intelligence is kind of the wrong word to use about it. So, I mean, I'm very sort, quite inspired to hear you say that actually. Do you think lots of other people share your view, or do you find creators generally are more on the worried side of the equation?

Nouf Aljowaysir: People who are familiar with how the technology works are not as worried, similar to me because they understand it. But I think others who don't understand it, there's just too much media hype around this and too much doom and gloom headline that are causing a lot of kind of ridiculous worry without any nuance being in the discussion there.

Kate Bevan: So how do we make people, particularly creators feel a bit more reassured about it, that there's a role for them in this?

Nouf Aljowaysir: Well, for one, I think projects like mine probably showing the failures of AI, showing where it has failed and showing that it's a process that you have to really train it, talk to it, show it more things, to really get the cool output. Even people who are using prompts to generate stuff, they go through a very long process to get that final result. So, the more that we talk about that process to actually get something looking nice from AI, the more people see, wow, it's not like snap of the fingers that simple to get something.

Kate Bevan: So, I'm going to come to my final question. Do you think AI is going to kill us?

Nouf Aljowaysir: No. Not at all. That's so ridiculous.

Kate Bevan: So, you are confident that that's just a scary narrative, right?

Nouf Aljowaysir: Yeah. It's a scary narrative borrowed from Hollywood movies.

Kate Bevan: I think also for me, it seems also to come from the Silicon Valley approach, which goes back to what we were talking about before about this particular paradigm of it where the Silicon Valley men are talking up the potential doom of it, partly to just make sure that they're organizing some of the regulation, they've got a stake in it.

Nouf Aljowaysir: Right. And I think they're profiting from this media hype. They're profiting. These movies get the most views and the most clicks. So, I think there's kind of this play between Silicon Valley and Hollywood that they're doing in to increase your interest and your fear around this technology. So, it's all really branding and marketing.

Kate Bevan: So, I suppose what you are doing, certainly for me, your film made me feel more reassured that it is very much a conversation and something that we all have a stake in.

Nouf Aljowaysir: Yes, totally.

Kate Bevan: Nouf, thank you very much indeed. You've been a fantastic guest. Thank you.

Nouf Aljowaysir: Thank you so much. Really enjoyed being here.

Kate Bevan: The AI Interrogator is an Infosys Knowledge Institute production in collaboration with Infosys Topaz. Be sure to follow us wherever you get your podcasts. And visit us on infosys.com/IKI.

The podcast was produced by Yulia Debari and Christine Calhoun. Dode Bigley is our audio engineer. I'm Kate Bevan of the Infosys Knowledge Institute. Keep learning, keep sharing.