AI, Art and New Technology | Threat or Opportunity
So for me, art is because I'm a conceptual artist. Art is just like the middle between things. So I make an art. It no longer belongs to me.
It belongs to the viewer. And it it's a middleman between the viewer and anything else. So in some ways, the artist can be an antenna, right? All art is, you know, between you and beauty or you and like, you know, whatever painting is that you're doing, whether you're trying to reach it, like, makes that it's connection, that's what it makes, The three artists that are going to hear from this evening has collectively shown their artworks at museums across the Western Hemisphere. The first of those three artists is Brett Amory.
Brett has sold out shows from London, England, to Los Angeles, California and New York City. Beyond the sold out gallery shows, Brett has shown his artwork at the National Gallery of Art in Edinburgh, Scotland, the National Portrait Gallery in London, England. And perhaps testament to the beauty of his creativity is he is without a doubt a painter's painter, something that is deeply admired by his peers for his technical acumen. With traditional oil paint. In the last handful of years, he's begun to incorporate new technologies into the traditional medium, using large language models and artificial generative intelligence to do so.
More recently, he was awarded the prestigious Jackson Pollock Lee Krasner Foundation grant here tonight to tell us about that confluence of his creativity and his artistic practice. Please welcome to the stage. We're just going into a round of applause. But Emery. The. Hello. Thank you for having me. Commonwealth Club.
And thank you, Hugh, for, gracious, for that nice introduction. So I want to start the talk. Talk about my history or, like, my relationship with with technology. I don't I'm kind of in between, like, a techno optimist and a dreamer.
I'm not quite a dreamer. And I'm, The more I use advanced technologies, the more concern I become. But I'm also very optimistic, and excited about, like, how accessible these tools are. I moved to San Francisco in the late 90s. My first job was Fedex, Kinko's or just Kinko's. Kinko's back then.
And when we would get slow at work, I would experiment with a copy machines. I moved out here to study. So I wasn't painting or making installations.
So when you get solo work, I would experiment with the copy machines and try to get them to do things that they weren't necessarily designed to do. Use them in unconventional ways. Try to kind of break them in some ways. That process led to using and experiment with Photoshop and and scanning my drawings and, and collage is where collages that make it Kinko's and experiment with it with with those those collages. That process led to paintings or drawing painting where I got my, undergrad and and painting and the the process of using Photoshop to compose digital composites for paintings was always, has always been part of my process.
All my work started out has, usually starts out with a digital image or some sort of composite or montage of some source, and then that image is translated into either into a drawing and then a painting or a study and, and a larger painting. In 2015, I came across the celebrity Gan. Don't know if you are familiar with that.
It's, in Goodfellow. It was inspired by in Goodfellas paper on, gender of adversarial networks and celebrity. Gan was basically a neural network that was generating images of people that that don't exist.
And when I saw that, it really blew me away. So from that point on, I really wanted to learn how to program these things. Now, I'm not a programmer, I don't code. And I quickly realized that that, that type of or that form of programing design and neural networks, machine learning is very sophisticated. Between 2015 and 2022, 2022, I took a couple classes in Python trying to learn how to program so I could try to learn how to design neural networks to use in my work. And I quickly would give up.
2021 I got covet, and during Covid, I was determined to learn how to program. So sign up for our bootcamp in Udemy in Python. And I was doing like three hours a day tutorials. Then after Covid was done with Covid, I did a ten day silent retreat. And when they lifted the veil of silence, that's when you get to meet everyone. And there are 25 men at this retreat.
25 men and women. And they separate the men from the women. And at the end of the retreat is when you get to talk and meet everyone. And I started talking to some of these men, and I would say 20 out of the 25 men were machine learning engineers.
And one of those engineers was, Alia Suk's of her, and I didn't know who he was at the time. And we were having conversations about, you know, programing and where it was going or machine learning where it was going. And he told me not to worry about learning how to code because everything was about to change. So I got out of that retreat in August of 2021.
And then in November, ChatGPT was released to the public. And the first thing that I did was how to generate code and teach me how to run that code to collage three images together. That was like the only parameters that I set. And these are some of the this is these are some of the images that I had ChatGPT write code to collage.
So when I got out of that retreat, the whole wave of text image generation and then just, I, generative art or generative AI started. These are some of the images that I started collage and or started using, text image generation platforms to play around with this was, like the the first idea I had using text image platforms was to generate images, to create some sort of disruption similar to what I was doing at Kinko's with the copy machines. So I generated a text prompt that make a lot of sense. There's some existential language buried in between, like a lot of nonsensical language, key symbols and, Unicode and a lot of gibberish. And I would take that prompt and generate thousands of images to see how the image would change over time without changing the problem.
So one of the, one of the first ideas was just to use the same prompt over and over and over on the same platform to see how the image would change, and try to create some sort of disruption within the model. And then I would take that same prompt and try to use it on on different platforms. So start out with crayon and then Dall-E and stable diffusion Midjourney, and use the same prompt across different platforms to see how the image would change based on the different data sets. And those are like some of the first parameters, or like some of the first ideas I was playing with, with, with image generation. And I wasn't doing anything other than generate an image of the images at that point and point an experiment and.
These are some of the example images that I generated initially, but just using that same prompt. And then after experiment for about three months generate I probably generate close to a hundred thousand images. I started taking some of those images, say like one out of 1000 images, and taking that into Photoshop and adding or subtracting from that image, and then outputting as either a screenprint or screenprint base layer and adding paint on top of it. So I started making objects from some of the selected images that fit, like a, some sort of narrative or theme that I was going for. And as like a, it was a somewhat of a back and forth process where I would take one image, taken Photoshop, add and subtract from it to it, and then output that a screenprint process, take a photograph of it and, upload the image back into Photoshop or some, generative AI platform and either expand it or add things to it and then output it again. And the second layer would be maybe, airbrush or oil, some, some sort of oil painting.
And I would start out with studies for smaller paintings or just would incorporate for color process screen printing to figure out how I was going to paint it larger. So these are some of the images, the initial images that I was painting from the text image generation images that I was generating, and this is the one on the left is a digital montage. And then the one on the right is a study which is a little bit smaller.
And then that study was used to figure out how to make a bigger painting. And then these are some of the this is the bigger painting that was generated from or made from the study in the digital collage. And again, same thing.
This is a digital montage. And then a study with a little bit of oil painting embellishment. And then the larger painting was made from that. And 20, 21, I was actually at the beginning of 2022, I was given I was added to the artist Beatles or the, Dali gave me a bunch of credits to try out their Dali 3.0, and I took that same prop that I have been using over and over and over, and I added more nonsensical language to it to try to create like more of a disruption.
Dali started started giving me these weird images that looked like alien advertisements with indecipherable texts. And it took me a little while to figure out, like or like to realize that this was the glitch that I was after looking for. So I started a body of work. The first body of work.
After using image generation for a couple of years within, first body, where was that was called avatars of our own making. It was about the disruption of communication and language. And there's right around the same time where, like, the hallucinations were starting to being talked about and how these limbs were hallucinating and, just doing things that they weren't designed to do.
And these are some of the other images. And I started realizing how using generative AI is was more than just using like, Photoshop or a tool became my most like a collaborator in some ways. And it was multimodal.
So I teach I was teaching AI and art AI in traditional art class. And one of the first things I tell my students is, is don't marry yourself to one application on one platform or one tool because as technology changes too fast is the exponential growth curve is just moving exponentially every week. So it's more important to come up with the workflow.
And like a this like multimodal, multimodal way of working. So I realized that using these tools was like a way to generate ideas and, you know, almost like this, like feedback loop was feeding back into itself. And, kind of creating or like recreating this stuff and generating new ideas from itself. And this is like more, more images from avatars of our own making. Last, last November, I had an exhibition in Oakland, and the entire exhibition started out with one text prompt. This is the image that was generated.
It was a black and white image. 512 by 512. And then I took that image and expanded it in Photoshop to be 168in by 130in, and then output that for color process screenprint in print for color process. And this was the backdrop for an installation that was in the gallery. The backdrop was in photographs and then uploaded back into Photoshop, where I added more elements to it.
So you can see some of the language and parts of it. So I digitally, added the original or the photograph that was taken of the the original backdrop screenprint, then added more things to it and then output it again. As for color process screen printing on top of it, and then photograph that imported it back into Photoshop and then added images or collage images for these paintings and some of the paintings that are coming off, a combination of oil painting and screenprint process. So this part here in the middle of the figure is oil painting, and the part on the right is screenprint for color process screen printing. And this is all oil painting. Right.
So at this point I've been using, I was had thousands of images I had generated a lot of these images had that indecipherable text and that image, that indecipherable language started giving me ideas, started to create something larger than a just a body of work of of images or installations and I started thinking about creating some sort of fictitious, religion based on, like, some of the stuff that I was generating. But then I realized I wasn't I'm not really religious and grew up studying religion, so I scrapped that idea. Then I was going to create some sort of like fictitious cult, but I don't really know much about cults, so I didn't do that. And then I started thinking about the way evolution and the way we've evolved or, you know, humans, the human species has evolved and we evolved through language. So I started, I started treating a, constructed language last January called I glyph at 913.
And it is turned into, like, this world building project. So now everything that all the work that I make kind of gets channeled through. I glyph 913 and, and it's turned into like this, like bigger project and a lot of the, a lot of the thematic structure behind it is dealing with like some of the ethical implications around the way in which, like effective celebration ism and transhumanism, just the speed at which these technologies are moving.
And it's a way for me to think about it and work through some of the, I guess some of the, angst I have around the rate at which technology is moving. At the same time, I was realizing like that. I could generate. Like, 100 images watching TV, right, watching a movie. And I realized that, I could continuously generate images and not really.
I wasn't doing anything with them. And this is it was really similar to, like scrolling and social media and is kind of giving me like this, like serotonin or a dopamine hit. And I started thinking about, like, fast fashion and, just like social media and it really caused me or I started thinking about slowing down and getting back to the handmade object.
So I started studying with this painter in Santa Cruz, and I was learning, I've been learning the Venetian glazing technique, which is of really started, I think. I think it was Titian came up with it in the 1500s or 16th century Baroque. And it's what Rembrandt used, and it's the apex of oil painting. So it's like very illusionistic and up, up, optical, so the idea was to incorporate that into the series and a way to slow down the process, but also merging with this, like this, like rapid pace of image generation.
And these are some of the this is one of the first images that utilizes that technique. And then boy rat, boy rat as part of I think 913 is like a construct within the language. So now the, the language and which is the backstory of the language, the origin story, determines like these philosophical constructs, for different bodies of work. And then this custom matters and the other body of work within norm and I go at 913 and it feels more like merging with technology and transhumanism and next ism and, this part here is utilizing that Venetian glazing technique and screenprint, outside. The make this Connacht tongue is another construct within Nam. And I looked at 913 and that's it.
Thank you. Okay. One more time for Brett. Thank you Brett. Our second presenter this evening is on a graphic honor. Earned a PhD in vision science from UC Berkeley.
She has dedicated more than two decades of research to the way that these new technologies, such as artificial intelligence, view the world that you and I collectively live in, as well as the real world implications of these technologies. Along the way, she's earned five patents and published more than 20 peer reviewed papers on the subject that research and information forms a foundation of inspiration for her visual art practice. And those artworks she has shown at New York City's Museum of Modern Art, as well as a contemporary Jewish museum and countless other institutions and museums across the United States. Perhaps the best testament to the beauty of her creativity is her list of collaborators. She's collaborated with everyone from composer of classical music Philip Glass, to multi-platinum Grammy Award winning artist Bjork, as well as, perhaps most beautifully, her husband, who is here in the audience with us this evening.
Here tonight to tell us about this confluence of research as well as a real world implications of these new technologies and her creativity. Please welcome to the stage with a round of applause, ladies and gentlemen, Honegger Sheikh. Thank you. You. Thank you. Here for the kind intro and for, organizing tonight.
So I put that word secret in my title, in part just to get your attention. But in part also because I felt like I had a secret. So, as you said, I'm trained in visual neuroscience and the science of AI vision. And during my research, I, I would frequently stumble upon views into the brain and views into an AIS visual processing unit that I thought were really interesting, powerful, beautiful, and weren't really widely known outside of a like small group of technical experts. So, they became the topic of my research and also my art practice, which is what I'm going to tell you about tonight. So I think you all know the five senses.
My specialty is the visual sense. And it's not just the AI, it's really the set of computations that happens in the brain to transform, light coming in into a, it's an algorithm into a visual percept. And there's an analogy, analogy with, AI vision. So the smartphone cameras, you have, or maybe it's a security camera, or maybe you took a Waymo. Here.
Or, medical imaging device, like they all have these cameras. And then there's an analogous algorithm that transforms, that signal into, like it, whether it be face recognition or the, technology underlying generative AI that, you know, that was talking about, that it's analogous between humans and AIS. And I'm really interested in some of the similarities between human and AI vision. So for one, they all they all use visual training data.
So that's coming from the visual environment, whether it comes from your eyes or from a set of photos on the internet or videos. It's a large volume of visual training data. And then from there, the purpose of learning is to basically condense that training data into something that can be reused. And so it's a relatively small number of what I'm calling visual primitives here. You can think of these as, what the neurons learn. And that's, that's where my research is in those visual primitives.
And those are useful because then they can be recombined in many complex ways, to form this really robust and flexible perceptual system. And I've kind of simplified it here. But the humans and I really share this common, learning approach with human learning. It's over the course of evolution as well as, during our lifetimes. Okay. So this was inspirational for me as a scientist.
As early as in the 1600s, it was like recommended for like a wealthy art patron who wanted to form an art collection to actually, not just put paintings in it and art, but also put, these collections of, like, scientific specimens, often naturalist specimens like this. This is a collection of butterflies in Vienna. And so the the specimens were part science, part entertainment. And, you know, meant to wonder.
And so I, I felt like I had a lot of specimens. They weren't naturalist specimens, but technological specimens. And so I started to gather them and think about ways of, displaying them and these, these patterns here I actually got from, inception, which is a 2014 model from Google. It's open source. So I can actually open it up and look inside. And it's just a jumble of numbers, but you can visualize the numbers and start to figure out what patterns the neurons learned.
For. And those are those visual primitives I was talking about. And so by creating this collection of technological specimens, I felt like I was, preserving an important moment of AI history. That 2014 time is when, these models really started taking off for the first time.
And this one is actually now hanging down the road at the new, UCLA building across in their technical technological LA lab. And so here's one that you can see a little closer up view. And I'm going to walk you through some examples of my artwork and, share some of the properties of these visual primitives. So the first property is that they're emergent.
Even though I designed the artwork, the patterns themselves, I didn't design. They emerged through the training algorithm, which I think is pretty cool because it's kind of, like I'm working collaboratively with them and I don't have full control over them. Another property is that these primitives are the optimal distillation of the training data, which basically just means that during the training, they optimally figured out how to divvy up the space. And each neuron was like, I'll take this corner, I'll take this corner. And then they distilled the photographs, down. So like, none of these patterns you would directly see in a photograph are you're not seeing corners of the photographs in these patterns.
Rather, you're seeing distillations of common patterns. That felt important to the learning algorithm. Okay. So this one, is a little more complex. And the patterns.
And so that's also a, property is that these AIS, are they're deep learning algorithm. So that means they have a lot of layers of neurons. And the early layers tend to be really simple. They focus on like and what they call edge detection. So it's like a black white edge.
And then later layers, become more and more complex until you get like face detectors. I was really excited when I hit on this layer. It's I label them all with the layers. And so this is called inception layer three A. And
I feel like it's my favorite layer because it, it's much more complex than a simple edge. There's a lot of richness to these patterns. And yet they're still abstract.
You can't quite put them into words and describe what they are. Here's another one, from the same series. Another way to think about this is that, together, a set of these would form a visual grammar. And so it's like a language.
It's, each one of them you could think of as a visual phoneme that is actually quite powerful because when you put them all together, like, this is the technology that would be underneath, like Google image search, for example. Okay. So you might be thinking, hey, that was an old model. I thought it was ten years old. So and technology has advanced quite a lot in the past ten years, especially this type of technology. So actually, yes, the technology has advanced, but the primitives are relatively similar.
Even in today's technology, they're not exactly the same. But conceptually that the gist of them is similar and especially at those early layers. And so what that points to is, that the training data, the training data, there's something universal in the training data that they're capturing, regardless of how advanced the technology gets, which I find pretty exciting and even more exciting is that these primitives are not just shared across AI models, but also humans and other mammals. Conceptually, they're very similar, especially those early layers. So that's not to say that humans and AI brains are the same. It's more pointing to, that they share a common learning, process and also somewhat similar training data.
And so all these learning systems are picking up on certain elements in the training data and the individual training data in the, in the world we live in, so that was also okay. Sorry. Now we're jumping to another project. So this is a series of photographs I took near my, West Berkeley studio. And at that time, I was thinking a lot about my postdoctoral research where I had been looking for, patterns in nature, actually, photos of nature and looking for certain elements that, seem to have worked their way into the primitives in human brains and other species. Brains.
And then also thinking about how those got into AIS. And, now AIS are looking at nature, so it kind of goes back to nature being the original training data. And in this series, I'm just kind of contemplating that cycle of it all connecting. And this is another one from that series, just again thinking about how like, I would see nature. Okay.
And then this one, I wanted to make a window where one could look through the window and have that experience of AI perception, like that initial moment of AI perception. So this is actually the first, convolutional layer. And this is this is a different model.
It's AlexNet. So it's actually came out in 2012. Excuse me. It was the first the first deep learning model to like pass this ImageNet. Challenge, which is a kind of well-known benchmark.
So it generated a lot of excitement. And I think you might be thinking, why are these patterns like, why are they so, they're so simple, but like I'm telling you, they're so powerful. And so, the visual task for humans in AI is actually, like, a really hard technical task. And so, the primitives are super helpful. Just for optimizing vision because they tell us, like, what to look for.
And so it's a shortcut for the brain or for a visual system. Here's another one I did in that series. And the flip side of being helpful is actually, bias. So we hear a lot about AI bias. And there's rightly some fears around it.
When I hear bias, I think about like my perceptual research and perceptual research bias has a very precise technical definition, which is basically the difference between, what we perceive and what's really there. And so the bigger the bias, the bigger that differences. And in my research, I had showed that these perceptual, visual primitives are basically a source of bias because they, they tell us, they tell us what to look for. So if you have an idea of what if you're out searching for something you tend to, your perception tends to go towards that. So another way to think about this window is the first moment of AI bias when you look through that window.
I've also made these mirrors now. And so it's a similar idea, but instead you can look at yourself through the mirror and like think about bias. Head on. And then this is just some recent experimentation I've been doing, working, with glass powders on glass to make glass windows.
And so I just wanted to wrap up with, summarizing a little bit. So these visual primitives I told you about, they're, they're both my scientific research area, but they're also my artistic material. And they're learned by brains and eyes because they're helpful. And they're they seem to be derived from, universal patterns in our visual environment. And, they're both helpful and they're biasing. So I know the theme of tonight's talk was threat or opportunity.
And so in that sense, they're both kind of a threat and opportunity. And in one, so there's my contact info if you want to, learn more. I also do open studios once a year and it just happens to be on Saturday. So if you're interested to come by, there's that information and I think that's it. Thanks. Thank you. Honor.
One more time for honor. Thank you. Thank you for education. The third and final presenter this evening is Sharif Wong.
£0.03 will be our final presenter before we go to the panel discussion. So if you were wanting to share a question to the community's collective curiosity, once we come to the panel discussion after three phase presentation and you haven't yet done so, now is a great time to do that. And we see there are question cards going around. If you haven't already shared your question, after we do the panel discussion, we're all going to go out the double doors behind you into the atrium and continue this very important civic discourse. So if your question does not get asked during the Q&A of the panel portion of this, please be the person who does ask that question when we are in the atrium.
Continuing the civic discourse over a glass of wine. At the same time, we will have the opportunity to meet no, no, not Thai. This is a realistic portrait drawing robot built on 60,000 lines of code written by Steve Bro. Sweeping along our third and final version of this evening is the founder of Icarus Salon. Aqua salon brings together creatives to focus on the ideas of art, activism and politics. Sorry Faye, for her part, has garnered numerous awards and foundational backing from the Berggruen Institute, the Rockefeller Foundation.
She's also going opportunities from Kid Lab at UC Berkeley, Creative Capital, and Mozilla Foundation. She also sits on the board at the Gray Area Foundation, which for more than 20 years has been the bedrock and cornerstone of the art and tech community here in San Francisco. Sorry, these projects are inspiring in the way that they bring together art and academic research into the realm of activism that is a big part of what she does with her current projects and beyond.
Here tonight to tell us about her creative practice, the idea of academic research and activism. Please welcome to the Stage with a round of applause. Sorry. They want. Thank you. Hi everyone.
Thank you so much for coming. I have notes because this is a very fresh talk. Okay. So, my name is Serife Wong and I'm a conceptual artist, so I don't really give talks about my work. I consider giving talks actually just part of my art practice.
I do a lot of different sort of pieces on websites or performance art. So for me, this is a performance art piece. And welcome to my talk artists. And I best not to skip therapy this week. Please go. Especially all those artists 50% right.
So, these are some of the questions I've gotten in the past eight years of my, practice giving these talks. Are artists real artists? Are this real artists? Yes, they are. And is I conscious? No. Well, I take over the world. No. Okay, great. We will move on. So let's talk about the real stuff here.
We are really talking about money and power because that's what I is all about artists. We know a lot about the AI market because it is very, very similar to the art market. Billionaires are there, right? And there's also a ton of people in the industry who are barely making ends meet.
And even when the product doesn't seem to have any practical use at all and not functional, people still believe it can change the world. Sam Altman here, who's giving a talk here? He's the CEO of OpenAI that makes Dall-E and ChatGPT. He is now looking for $7 trillion of investment in AI. Now I'm guy guy.
Yeah. You know, we how much is that? How much is that? All right. So, artists, right. We are a talented bunch of hustlers. Still, we just sold a banana that is duct tape to a wall for $6.2 million. And the buyer of it, he ate it, right? He ate it, you guys, he just ate it.
But you cannot eat a million bananas. The math does not work out, just like I. So the AI market is more bananas than the art market.
So if it's not our jobs, that's worth 7 trillion bucks. What is, AI tools can be used in creative ways, but the creative industries are peanuts. Next to the big contracts from the military, Microsoft is one of the largest military contractors in the United States. The art stuff is a trick. How do I know it's a trick? I recognize them because I am married to a magician. Brad Barton, reality thief available for Christmas parties Brad is a mentalist, which means he can read minds and he makes predictions so he pretty much can do what tech companies want us to think I can do.
But there's a really big difference. At the magic show, you agree to get fooled, the magician knows you and he shows you what he wants you to see. What am I marketing? There is no disclaimer.
Companies direct your attention to the magic of instant art and poems, and they're hiding their tricks. Generative AI is a tool of power masquerading as a tool of knowledge. And this marketing, it had better work. There's a lot of money riding on it. We have got to buy those products. We need smartphones, smart watches, smart homes, smart cities.
Every time you see the word smart, replace it with the word surveillance because that what is what AI is. Oh, man. Therapy. Right. So AI is a tool, but it's not merely a tool for creativity. It's a tool for control.
AI is a vast surveillance infrastructure. ChatGPT works because it's trained on data. Every time you use it, it collects your data, everything you put in. But even that is not enough. Companies need more data. Data takes up space, it uses up energy, and it's bundled together with buildings, cable, satellites, methods and institutions.
You can buy say, and anybody, right? You can buy and sell data, stockpile and use data. So data centers are devouring power. Already 32% of Ireland's electricity goes to data centers. That is the next milestone here in the United States. And so please think of California's grid.
Think of our fires. Big Tech's carbon emissions are soaring 30% from AI. Their solution is to buy and invest in nuclear power. That's where they're at. And who's really going to pay for all that? Right? It's the poor, the marginalized.
The land is, you know, the land that they're putting datacenters on. That's not right here in our backyard, that's over there, right somewhere else. And also labor is exploited in these systems.
People are training these tools and cleaning them up all over the world, getting paid like a dollar an hour. People are being processed through systems using facial recognition. The governments control, deport, imprison and even kill people using AI. This isn't about technology and art.
It's kind of about who is going to get to survive. I really love this painting. It's by Walton for the fallen biology extinction of the passenger pigeon, and it hosts some of my concerns that I have with AI.
My main one is that power is getting concentrated. Only a handful of companies have the wealth and compute power to build these systems. And then monopoly, like the monopoly they've had over the internet. It's not just economic, it's a way to shift power to whoever gets to hold those strings. Look at Elon Musk and Trump right now.
All of these concerns culminate towards a troubling direction. Everything made visible to machines becomes easier to control. That's an infrastructure of authoritarianism, whether wielded by states or companies. They're not separate problems. It's the system working as it was designed.
We have to think of AI in this context. AI is situated in a political economy, and it's in service to big tech companies and not in service to the public interest. Right? So what should artists do with that knowledge? Whose power should we uphold? You know, I'm criticizing AI because I really love AI. Blade runner is like my favorite movie. I've loved AI. We've all love AI. We've all seen those movies.
It clouds our judgment, you know, in some ways, because we automatically think it's so smart and we love it. And I love technology. I want it to be used well. I want artists to code, build, create. Whether they use AI or not. But we have to see through the marketing art. Washing is old.
Powerful interests have always used art to distract us from what they don't want us to see. Are we using the tools or are the tools using us? Are this know how to see? Both Art and I are in service to a vision. Do we want to serve the vision of the already rich and powerful, or do we want to claim the power of the arts as our own? These are, some recommendations I have. If you want, you can take out your phone and take a picture. Ruha Benjamin Dan McQuillan, Cathy O'Neil, Timnit Gebru, and Khadija Rahman's Logic magazine.
They're great sources for information to learn about the concentration of power. And I also have my own, public, education forum. I'm doing performance art on TikTok and Instagram, trying to reach young people. But I have this website, Artificial Life Coach, and it has a whole bunch of resources for everyone. It has artist activists, papers. It makes I really accessible for people.
And you can learn about some of the impacts they have as well. All right. Thank you. And now I'd like to invite all three of the artists to join me, for our panel discussion this evening. This is where we will focus on your questions that have been shared with Gerald Harris, who's now walking around. If you can get him your question, and we will focus on those in just a moment here.
So tapping into the collective curiosity of our community, as mentioned, if we don't get to your question, please be the person who does ask that question at a time. That civic discourse is more important now than ever before on these topics. That will not just change the art world, but will change society as a whole. We will do that in just a moment.
When we go out beyond the double doors and consider the civic discourse over a glass of wine, as well as meeting no, not tie the realistic portrait drawing robot. So I'm going to have a seat and we're going to start focusing on your questions here. So you mentioned the perspective is is probably very rightfully, cynical. I think that it seems like on one hand, there is a fear.
Of. Displacement of the artist. And you take a more macro perspective on this beyond just the arts and, and almost a societal perspective, which I think is very justified. Do you see some form of where you think like, oh my goodness, this might be a very positive for, for the arts, for artist. Is there something that, might be the opposite end of the spectrum from what you just shared with us? Yeah, I mean, I do have a positive example because, I don't like big tech companies kind of meddling in the arts, but AI systems don't necessarily have to come from those, places.
Right. There is a really great little group called Kahuku Media in New Zealand, and it's, a Maori communications media group. And, they were pitched to give their language, the Maori language, up to, company. That was working for OpenAI. But to train ChatGPT on and they said no because they compared it to sort of colonialism. Right. They're like, you want to take our language now for $20 for two hours, and then you'll sell it back to us later.
So instead, you know, there was a computer scientist working there and he like, googled and he taught himself how to do just like you. He, like, learned how to do some AI. And he made their own system, and they trained their own system for language. So all those kind of rewards of like the like preserving indigenous languages is really, really important.
And young people are on computers, they're on to, you know, they're not talking amongst themselves. And people are all over the world. Right? So technology can bring people closer together and AI systems can help preserve and help people communicate. Right. But it needs to be bottom up. And that's when I'm really supportive of it.
Thank you. I want to go and turn our attention and thank you, George, to your perspective. But you seem to take, a unique, almost contrarian perspective on this, where it seems like a lot of artists, particularly painters and people that are, really knee deep in traditional media, filling, feeling a sense of fear around the potential that it's going to infringe on them. But with with your artwork, you really seem to find a place of empowerment with putting it into your painting process, and it's almost become very informative, inspiring for you. How do you see this
dichotomy of good versus bad displacement versus empowerment? I think it's important to have the understanding that these tools are being marketed to to us as a make things more convenient. I know that my approach using these tools is more of, they're sort of more and more more than tools for me. Like they're these like collaborators, but like, I think you have to be critical of, like, what they offer. I don't think they're necessarily going to replace the artists and support and have like, a vision and a voice, and not rely solely on the technology to generate the work. I think I come from a unique perspective where I had a practice for 25 years before using, you know, advanced technology.
I mean, technology has always been a part of the process, but not like generative AI is, I've gotten into a lot of conversations with, like, traditional artists, about using these tools. And it's like, it's not like talking about politics. You know, like there's a lot of pushback and resistance.
And I can I understand that pushback and I can see why people are scared. And don't want to use it. I think my approach is unique because I've used different things like Photoshop. You know, that's what actually got me into traditional object making. But I think it is important to be critical.
The more I use it, the more critical I become in this project, because the more research I do and the more I learn about the how technology is being marketed, taught to us. And like, Yeah, I mean, I think it is something to take into consideration, and use of use it with, I don't know, just be be mindful of that. You know. Thank you. So there's a couple of these that have a bit of overlap. And I think to kind of co-lead a couple of these together speaks directly to something that your research goes to on.
And the idea around the real world implications of this, of using these technologies, if we are to go ahead in the future, ten years, looking back on on today, what what would be the warning to ourselves of like, oh my God, what are the real world implications of for the the kids getting on there and using these things? Or if it's just for artists that in agency as the creative process, what do you think make of that. Right. I mean, I think what Brett was saying about artists having their own voice is really essential because, you know, that's what I'm attracted to in art is when I see the artist's message and voice and their hand in the work. Right. And so as these tools, I mean, as they're marketed to us as artists, like, we have to figure out, be creative and figure out ways to keep our voice in that process. And, you know, and maybe it's getting more challenging to do that.
But the constraints also, like, we can live up to them and be creative and figure out new ways to deal with them. And I think artists have done that. You know, forever in different ways in a sense. So we just have to rise to the occasion of like showing our, our human humanity in the face of increasingly powerful technology. Thank you. Generative AI uses stolen work from smaller artist uncredited.
On top of the environmental repercussions, how would you justify the use of artificial generative intelligence? Despite these issues? What do you think, Sheriff? Yeah, I feel like this, you know, comes back to, some of the education that you were just sharing with us, which was. Which was wonderful. What do you think about this? Hey, can you repeat the questions? Absolutely.
Generative AI uses stolen work from smaller artists uncredited, on top of the environmental repercussions. How can we justify the use of generative AI despite these issues? Yeah. Second time didn't help as much. Is that's a really good question because, the training material is stuff. Now, copyright stuff is happening, right? There's a big, lawsuit at the New York Times with, open AI because they trained without permission.
But it's millions of millions of, every website, every art is just been completely vacuumed up. And so copyright law doesn't really protect against that. And, but it's still a sort of theft because it's a theft of labor in time. A lot of artists have seen their stuff on there. And, and it's, you know, it's heartbreaking when you've worked so hard, you're not making any money. And then suddenly people can just, like, replace you, right? So on the ethical side of that, it's difficult to say that it's okay to use it.
Right. But, individual use and putting it upon an individual person and the choices that they make is also very unethical. But it's not about an individual, using the tool. Like, I don't feel bad when I use AI tools, right? I use them a lot because I'm often, testing them and seeing where they break and seeing what their problem is.
Right. And, I do get like some like, oh, this is bad for the environment kind of feelings, but, it's not about our individual use cases or something to make us feel like it's up to us, but it isn't. It's about a collective sort of pushing back. It's about those policies. Kind of adding on to that here. Can you talk about the relationship between art and human communication in your artworks? How does it speak to the people? What do you think about that? Brett.
I don't really know. I mean the work I'm making now is not as accessible as the work I was making before, and I think it's partly because I'm still figuring out what that thing is. I, I mean, I've always been of the camp to make it accessible for everyone and that's well, I mean, that's well, that's why I think street art is so good because it's like you don't need a PhD or, you know, MFA to understand what the artist is talking about or trying to communicate. I mean, I think powerful art can grab you visually and conceptually, intellectually.
I think the power of metaphor is important, especially like the stuff I'm doing right now, like I'm trying to communicate and certain things that, that concern me. But using a made up language to communicate these ideas without hit people over the head of like, these ethical implications of what the technology could potentially cause or, where it could potentially go, I've kind of, both camps of, I've, I want the work to grab me visually and I wanted to speak, I want the work to speak for itself, but also like when, when the viewer has to work a little bit to, you know, like get to the truth of the work. So, yeah, I don't I'm not really quite sure. I mean, I would say right now where I'm at with my own work is, yeah, this is definitely communicating certain ideas, but I think the viewer has to do a little work to get to the to the I don't know what's behind the curtain. Yeah.
But again, you know, this is what I'm doing right now. It's relatively new, and I think it's going to take some time for me to like, unravel, like and simplify and make it more accessible. So I don't know if that answer your question.
Absolutely. Let's take this back a step. Who owns the copyright when you create art and say Midjourney or Dall-E or beyond, whose art is it? AI art is not copyrightable. I kind of think I mean, I kind of relate to like postmodernism, you know, where everything was appropriated. Readymades.
I mean hip hop, you know, like in the 90s and then 2000 where everything was just remix, remix culture. I mean, I had work when I was doing another body of work, I had this artists, steal my work or, like, copy, copy my paintings and, and exhibit my paintings in galleries, and that sucks. You can, you can. My work has been scraped like like 1F1. The image text image generation started. You could, you could put in a Brett Emery waiting painting that would generate an AI image that looked like a bread emulating painting.
But I'm kind of like, the way I think about it is it's up to the artist to reinterpret and on it and like, it's not. I mean, I think it sucks that, like, artists, work's being stolen, but that's not that's not a new thing. I think it's been happened for a long time. Let's add on to that that you take us on a really interesting road here. So that kind of tapping into this next question here in the question cards, Marshall McLuhan, 1964 writes that or says rather in a debate, you know, that I see that great art is an early, distant warning system. It's used to warn the old culture what is about to happen to it because of the new culture, and that art is preparing the old culture for the psychosocial impacts that are about to take place.
If Marshall McLuhan is still holds true today, that art is this early warning system, what are the three of you? What are your artworks warning us? What are the psychosocial impacts? I mean, I'm trying to, Create something that, maybe doesn't, doesn't explain itself, but the power behind it is the concept behind it that I'm putting into it is kind of about this connection between AI and art, or AI in the brain, in nature. And, that doesn't really change. I mean, to me, that feels almost timeless. And I, searching for those timeless qualities, of course, is very timely also with AI coming out.
But I'm trying to find those timeless qualities, that exist this sort of like primal qualities. I think for myself as like, I really don't know if I have a lot of things that you talked about in your presentation. About, like, I grew up in Virginia, skateboarding, punk rock. A lot of the friends that I grew up with are libertarians. There was never left. And, you know, like when Trump was first elected to the first time around, we would have conversations and we would clash.
And I thought it was it was disturbing that, like, we couldn't even talk about politics in a civil way without getting some sort of spoken emotional. And I started thinking about like, the way the country is being divided and like the ways, you know, surveillance capitalism and like, algorithms and like, manipulation of emotion, all these things are playing into it. And I couldn't help but think that there's something happening, you know, in the middle where if you get both sides fighting each other, then we're distracted enough, and it's the power of manipulation and magic. To implement more, you know, like merger control or whatever, you know, so, like some of the things that I've been thinking more and more about is like what Trevor Paglen has been talking about lately, like Psyop capitalism, you know, like using advanced technologies for persuasion, manipulation into categories and heard and, so I don't know if I mean, some, some of the work I've been thinking about and like starting to kind of play with this is consider in some of these, like, you know, power dynamics or, you know, manipulative sort of like structures that are being implemented. But yeah. So some of those ideas are we could pull that into what you said during your presentation, which I thought was great, is every time you hear smart, replace it with the word surveillance, smart home, smart phone, surveillance phones.
Fancy. How can you add on to to some of what Brett saying and, you know, to give a bit of context to Trevor Peyton's work, this idea of, you know, we we were talking earlier about the idea Silicon Valley in some ways becomes effectively a proxy for social engineering. They have so much data on so many people.
How do you think about this with your work and some of what your current project is? Yeah. So, you guys want to hear my theory of art? Okay. Okay, great. Well, onward.
So for me, art is because I'm a conceptual artist. Art is just like the middle between things. So I make an art. It no longer belongs to me. It belongs to the viewer.
And it it's a middleman between the viewer and anything else. So in some ways, the artist can be an antenna, right? All art is, you know, between you and beauty or you and like, you know, whatever painting is that you're doing, whether you're trying to reach it, like, makes that it's connection, that's what it makes, you know. And so in my work where I'm being I'm being critical. I'm trying I'm not trying to just tell people stuff.
I, I'm trying to move them with some emotion. You know, they can see that I care you, you know, I want to connect to people. I want them to experience how I feel about AI and come with me on that journey and try this change, shift their lens a little bit, you know, and then I also want to view and hear back. Right? I want to be that middleman between something and something else. Because art throughout history has always meant so many different things, and it's always been so expansive, like the history of art. It's just like, oh, it's only paintings.
Oh, actually, you know, we're going to performance art, we're going do video. Art's just growing and growing and expansive, and I find I and what Silicon Valley is doing very reductive, very limiting. That art should look like this. It should be on the computer. And it's going to, you know, be only what we've trained on.
Most of art does not exist on the internet. You know, it's all over. Most of our language, what we say, what we've written, it's not all on the internet. None of that is captured. So when we reproduce stuff with AI, we are constantly limiting and reducing, you know, making it less. And I think art we have like artists, we have a duty to be this expansive thing.
You know, art can be a thing that is a moment of healing. It is one of the most powerful things in the world, you know, because like a lot of fear is running around with AI going to take our jobs, is it going to, like, ruin the world? All sorts of things. And art is much more powerful than that fear. Because art, when someone sees art, they get moved by it and they love it. That is a much more powerful emotion than fear.
You know, fear you might kill somebody because they're afraid, but with you love, you know, you could run into a burning building and save them, you know? So that's like what I think is a good counterpoint to Silicon Valley is the work of artists, which has never, you know, which is always constantly very, very important. And it's extremely crucial now in this time that we're living in like a turducken of, of like crises, right? A pandemic, climate change, politics all around the World war. We have a response.
We have a duty like we need to survive together. Thank you. And we got time for one more question here. At the start of the evening, I ask everyone in the audience here.
You raise your hand if you identify as an artist or arts professional and 40 plus percent of the hands go up, what is, you know, coming back to something that that you said in in your talk and I mentioned in introducing you know, is this idea of collaboration and the idea of reaching across, you know, with Bjork, right? A musician with your husband, someone in your house. So on this idea of collaboration becomes profound and you mentioned it as well. Sharif had this idea of of being a communicative tool for, with AI and so on. And so night night shade or glazing. So to give a bit of context to the audience here, the idea of night theater glazing is this idea of almost like a defense mechanism of something that can be installed into the images that you create of you.
You take a photo of your painting, you can install this in there so that if it tries to be scraped and used by an, artificial intelligence model, that it will not function properly. Effectively. What do you say to to the 40 plus percent of the creators here this evening, that how should they look at this should be empowered? Should they put their head in the sand? Should they, reach out? What do you say for take action to people. You're asking me about? Absolutely. About this night night shade algorithm.
I've never used it, but, I'm not trying to. I don't think an AI algorithm would be very interested in what I'm making, you know, I think it really is very personal choice. Right? It depends on, what you're trying to protect and why. And, to what extent you want to do it. But I think it's reasonable to, you know, want to, think about it. At least think seriously about, like, how much you want to just freely offer your work, you know, your imagery, as you said, to connect people and put it out there like your the and the tenor of the artist versus, a more locked down approach.
Yeah, but I personally don't do anything. Perhaps. Brett, can you add on to that? Me? You're someone who use, like you said, 25 years of traditional art painting practice, and here you are, you know, experimenting with this new media. What do you tell, artists here in the audience the idea of how to think about either protecting their work from being scraped or should they dive right in? Well, how do you think about that? I mean, I think if you're really concerned about your images or your work being taken, then. Yeah, maybe protected.
Do you? I don't know, because. I think if someone's going to steal your work, then they're always going to be a step behind. You know, like, a a biter is always going to be a biter. Hop is always me a copy or so. Like, it doesn't really concern me. If someone wants to steal my work, then that's fine.
I'm always a step ahead. I encourage artists to embrace these technologies because they're not going anywhere. I mean, I think it's important to be critical.
But being that these technology is so accessible now, where you don't really necessarily have to have a degree in programing or CSS, agree to use them. Enables your, you know, scaling of the practice, your practice. Like my work has changed more in the last three years and it has in 20 years. And I'm doing like all these projects that outside of just painting that I've always wanted to do, but I couldn't do because I had I was limited because of, like, not knowing how to do certain things. I feel like every month something new comes out that, like, I can utilize for a different aspect of a project that I couldn't do before. So that scaling aspect of like my own practice and like, this sort of it's just the way like this, this language project is turned into like a world building project, and it's turned into something much bigger than I ever anticipated in the initial anticipated is, I think it's something that like, is only possible because these technologies are accessible.
So, yeah. Thank you. Thank you for n
2024-12-16 15:40
