Collaging, remixing, sampling—art always has been more than the sum of its parts, a synthesis of elements and ideas that produces something new and thought-provoking. Technology has enabled and advanced this enormously, letting us access and manipulate information and images in ways that would’ve been unimaginable just a few decades ago.

Privacy info. This embed will serve content from

Listen on Spotify Podcasts Badge Listen on Apple Podcasts Badge  Subscribe via RSS badge

(You can also find this episode on the Internet Archive and on YouTube.)

For Nettrice Gaskins, this is an essential part of the African American experience: The ability to take whatever is at hand—from food to clothes to music to visual art—and combine it with life experience to adapt it into something new and original. She joins EFF’s Cindy Cohn and Jason Kelley to discuss how she takes this approach in applying artificial intelligence to her own artwork, expanding the boundaries of Black artistic thought.  

In this episode you’ll learn about: 

  • Why making art with AI is about much more than just typing a prompt and hitting a button 
  • How hip-hop music and culture was an early example of technology changing the state of Black art 
  • Why the concept of fair use in intellectual property law is crucial to the artistic process 
  • How biases in machine learning training data can affect art 
  • Why new tools can never replace the mind of a live, experienced artist 

Dr. Nettrice R. Gaskins is a digital artist, academic, cultural critic, and advocate of STEAM (science, technology, engineering, arts, and math) fields whose work she explores "techno-vernacular creativity" and Afrofuturism. She teaches, writes, "fabs,” and makes art using algorithms and machine learning. She has taught multimedia, visual art, and computer science with high school students, and now is assistant director of the Lesley STEAM Learning Lab at Lesley University.  She was a 2021 Ford Global Fellow, serves as an advisory board member for the School of Literature, Media, and Communication at Georgia Tech, and is the author of “Techno-Vernacular Creativity and Innovation” (2021). She earned a BFA in Computer Graphics with honors from Pratt Institute in 1992; an MFA in Art and Technology from the School of the Art Institute of Chicago in 1994; and a doctorate in Digital Media from Georgia Tech in 2014.  


What do you think of “How to Fix the Internet?” Share your feedback here.


I just think we have a need to remix, to combine, and that's where a lot of our innovation comes from, our ability to take things that we have access to. And rather than see it as a deficit, I see it as an asset because it produces something beautiful a lot of the times. Something that is really done for functional reasons or for practical reasons, or utilitarian reasons is actually something very beautiful, or something that takes it beyond what it was initially intended to be.

That's Nettrice Gaskins. She’s a professor, a cultural critic and a digital artist who has been using algorithms and generative AI as a part of her artistic practice for years.

I’m Cindy Cohn - executive director of the Electronic Frontier Foundation.

And I’m Jason Kelley - EFF’s Activism Director. This is our podcast series How to Fix the Internet.

On this show, we’re trying to fix the internet – or at least trying to envision what the world could look like if we get things right online. At EFF we spend a lot of time pointing out the way things could go wrong – and jumping in to the fray when they DO go wrong. But this show is about envisioning, and hopefully helping create, a better future.

Our guest today is Nettrice Gaskins. She’s the assistant director of the Lesley STEAM learning lab at Lesley University and the author of Techno-Vernacular Creativity and Innovation. Her artwork has been featured by the Smithsonian, among many other institutions.

Nettrice has spoken about how her work creating art using generative AI prompts is directly related to remix culture and hip hop and collage. There’s a rich tradition of remixing to create new artworks that can be more than the sum of their parts, and – at least the way that Nettrice uses it – generative AI is another tool that can facilitate this kind of art. So we wanted to start the conversation there.

Even before hip hop, even the food we ate, um, poor people didn't have access to, you know, ham or certain things. So they used the intestines of a pig and then they created gumbo, because they had a little bit of this and a little bit of that and they found really creative and innovative ways to put it all together that is now seen as a thing to have, or have tried. So I think, you know, when you have around the world, not just in the United States, but even in places that are underserved or disenfranchised you have this, still, need to create, and to even innovate.

And I think a lot of the history of African Americans, for example, in the United States, they weren't permitted to have their own languages. But they found ways to embed it in language anyway. They found ways to embed it in the music.

So I think along the way, this idea of what we now know as remixing or sampling or collage has been there all along and this is just one other way.  I think that once you explain how generative AI works to people who are familiar with remixing and all this thing in the history, it clicks in many ways.
Because it starts to make sense that it is instead of, you know, 20 different magazines I can cut images out and make a collage with, now we're talking about thousands of different, pieces of information and data that can inform how an image is created and that it's a prediction and that we can create all these different predictions. It sounds a lot like what happens when we were looking at a bunch of ingredients in the house and realizing we had to make something from nothing and we made gumbo.

And that gumbo can take many different forms. There's a gumbo in this particular area of the country, then there's gumbo in this particular community, and they all have the same idea, but the output, the taste, the ingredients are different. And I think that when you place generative AI in that space, you're talking about a continuum. And that's kind of how I treat it when I'm working with gen AI.

I think that's so smart. And the piece of that that's important that's kind of inherent in the way you're talking about it, is that the person doing the mixing, right? The chef, right, is the one who who does the choices and who's the chef matters, right?

And also, you know, when they did collage, there's no attribution. So if you look at a Picasso work that's done collage, he didn't, you know, all the papers, newspapers that he took from, there's no list of what magazines those images came from, and you could have hundreds to 50 to four different references, and they created fair use kind of around stuff like that to protect, you know, works that are like, you know, collage or stuff from modern art.

And we're in a situation where those sources are now quadrupled, it's not even that, it's like, you know, how many times, as opposed to when we were just using paper, or photographs.

We can't look at it the same because the technology is not the same, however, some of the same ideas can apply. Anybody can do collage, but what makes collage stand out is the power of the image once it's all done. And in some cases people don't want to care about that, they just want to make collage. They don't care, they're a kid and they just want to make paper and put it together, make a greeting card and give it to mom.

Other people make some serious work, sometimes very detailed using collage, and that's just paper, we're not even talking about digital collage, or the ways we use Adobe Photoshop to layer images and create digital collages, and now Photoshop's considered to be an AI generator as well. SoI think that if we look in the whole continuum of modern art, and we look at this need to curate abstractions from things from life.

And, you know, Picasso was looking at African art, there's a way in which they abstracted that he pulled it into cubism, him and many other artists of his time. And then other artists looked at Picasso and then they took it to whatever level they took it to. But I think we don't see the continuum. We often just go by the tool or go by the process and not realize that this is really an extension of what we've done before. Which is how I view gen AI. And the way that I use it is oftentimes not just hitting a button or even just cutting and pasting. It is a real thoughtful process about ideas and iteration and a different type of collage.

I do think that this bridges over into, you know, an area where EFF does a lot of work, right, which is really making sure we have a robust Fair Use doctrine that doesn't get stuck in one technology, but really can grow because, you know we definitely had a problem with hip hop where the, kind of, over-copyright enforcement really, I think, put a damper on a lot of stuff that was going on early on.

I don't actually think it serves artists either, that we have to look elsewhere as a way to try to make sure that we're getting artists paid rather than trying to control each piece and make sure that there's a monetization scheme that's based upon the individual pieces. I don't know if you agree, but that's how I think about it.

Yeah, and I, you know, just like we can't look at collage traditionally and then look at gen AI as exactly the same. There's some principles and concepts around that I think they're very similar, but, you know, there's just more data. This is much more involved than just cutting and pasting on canvas board or whatever, that we're doing now.

You know, I grew up with hip hop, hip hop is 50 this year, I'm 53, so I was three, so hip hop is my whole life. You know, from the very beginning to, to now. And I've also had some education or some training in sampling. So I had a friend who was producing demos for, and I would sit there all night and watch him splice up, you know, different sounds. And eventually I learned how to do it myself. So I know the nature of that. I even spliced up sampled musics further to create new compositions with that.

And so I'm very much aware of that process and how it connects even from the visual arts side, which is mostly what I am as a visual artist, of being able to splice up and, and do all that. And I was doing that in 1992.


I was trying to do it in 1987, when the first time I used Amiga and DePaint, I was trying to make collages then in addition to what I was doing in my visual arts classes outside of that. So I've always been interested in this idea, but if you look at the history of even the music, these were poor kids living in the Bronx. These were poor kids and they couldn't afford all the other things, the other kids who were well off, so they would go to the trash bins and take equipment and re-engineer it and come up with stuff that now DJs around the world are using. That people around the world are doing, but they didn't have, so they had to be innovative. They had to think outside the box. And they had to use – they weren't musicians. They didn't have access to instruments, but they did have access to was records. And they had access to, you know, discarded electronics and they were able to figure out a way to stretch out a rhythm so that people could dance to it.

They had the ability to layer sounds so that there was no gap between one album and the next, so they could continue that continuous play so that the party kept going. They found ways to do that. They didn't go to a store and buy anything that made that happen. They made it happen by tinkering and doing all kinds of things with the equipment that they had access to, which is from the garbage.

Yeah, absolutely. I mean, Grandmaster Flash and the creation of the crossfader and a lot of actual, kind of, old school hardware development, right, came out of that desire and that recognition that you could take these old records and cut them up, right? Pull the, pull the breaks and, and play them over and over again. And I just think that it's pulling on something very universal. Definitely based upon the fact that a lot of these kids didn't have access to formal instruments and formal training, but also just finding a way to make that music, make that party still go despite that, there's just something beautiful about that.

And I guess I'm, I'm hoping, you know, AI is quite a different context at this point, and certainly it takes a lot of money to build these models. But I'm kind of interested in whether you think we're headed towards a future where these foundational models or the generative AI models are ubiquitous and we'll start to see the kids of the future picking them up and building new things out of them.

I think they could do it now. I think that with the right situation where they could set up a training model and figure out what data they wanted to go into the model and then use that model and build it over time. I just think that it's the time and the space, just like the time and the space that people had to create hip hop, right?

The time and the space to get in a circle and perform together or get into a room and have a function or party. I think that it was the time. And I think that, we just need that moment in this space to be able to produce something else that's more culturally relevant than just something that's corporate.
And I think my experiences as an artist, as someone who grew up around hip-hop all my life, some of the people that I know personally are pioneers in that space of hip-hop. But also, I don't even stay in hip-hop. You know, I was talking about sashiko, man, that's a Japanese hand-stitching technique that I'm applying, remixing to. And for me to do that with Japanese people, you know, and then their first concern was that I didn't know enough about the sashiko to be going there. And then when I showed them what I knew, they were shocked. Like, when I go into, I go deep in. And so they were very like, Oh, okay. No, she knows.

Sashiko is a perfect example. If you don't know about sashiko embroidery and hand stitching, there were poor people and they wanted to stretch out the fabrics and the clothing for longer because they were poor. So they figure out ways to create these intricate stitching patterns that reinforced the fabric so that it would last longer because they were poor. And then they would do patches, like patchwork quilts and they it was both a quilting and embroidery technique for poor people, once again, using what they had.

When we think about gumbo, here's another situation of people who didn't have access to fancy clothing or fancy textiles, but found a way. And then the work that they did was beautiful. Aesthetically, it was utilitarian in terms of why they did it. But now we have this entire cultural art form that comes out of that, that's beautiful.

And I think that's kind of what has happened along the way. You know, we are, just like there are gatekeepers in the art world so the Picassos get in, but not necessarily. You know, I think about Romare Bearden, who did get into some of the museums and things. But most people, they know of Picasso, but they don't know about Romare Bearden who decided to use collage to represent black life.

But I also feel like, we talk about equity, and we talk about who gets in, who has the keys. Where the same thing occurs in generative AI. Or just AI in general, I don't know, the New York Times had an article recently listed all the AI pioneers and no women were involved, it was just men. And then so it was a Medium article, here were 13, 15 women you could have had in your list. Once again, we see it again, where people are saying who holds the keys. These are the people that hold the keys. And in some cases, it's based on what academic institution you're at.

So again, who holds the keys? Even in the women who are listed. MITs, and the Stanfords, and somewhere out there, there's an AI innovator who isn't in any of those institutions, but is doing some cool things within a certain niche, you know, so we don't hear those stories, but there's not even opening to explore that, that person who wrote and just included those men didn't even think about women, didn't even think about the other possibilities of who might be innovating in space.

And so we continue to have this year in and year out every time there's a new change in our landscape, we still have the same kinds of historical omissions that have been going on for many years.

Could we lift up some of the work that you have, have been doing and talk about like the specific process or processes that you've used? How do you actually use this? 'Cause I think a lot of people probably that listen, just know that you can go to a website and type in a prompt and get an image, and they don't know about, like, training it, how you can do that yourself and how you've done it. So I'm wondering if you could talk a little bit about your specific process.

So, I think, you know, people were saying, especially maybe two years ago, that my color scheme was unusually advanced for just using Gen AI. Well, I took two semesters of mandatory color theory in college.

So I had color theory training long before this stuff popped up. I was a computer graphics major, but I still had to take those classes. And so, yeah, my sense of color theory and color science is going to be strong because I had to do that every day as a freshman. And so that will show up.

I've had to take drawing, I've had to take painting. And a lot of those concepts that I learned as an art student go into my prompts. So that's one part of it. I'm using colors. I know the compliment. I know the split compliments.

I know the interactions between two colors that came from training, from education, of being in the classroom with a teacher or professor, but also, like one of my favorite books is Cane by an author named Jean Toomer. He only wrote one book, but it's a series of short stories. I love it. It's so visual. The way he writes is so visual. So I started reinterpreting certain aspects of some of my favorite stories from that book.

And then I started interpreting some of those words and things and concepts and ideas in a way that I think the AI can understand, the generator can understand.

So another example would be Maya Angelou's Phenomenal Woman. There's this part of the poem that talks about oil wells and how, you know, one of the lines. So when I generated my interpretation of that part of the poem, the oil wells weren't there, so I just extended using, in the same generator, my frame and set oil wells and drew a box: In this area of my image, I want you to generate oil wells.

And then I post it and people have this reaction, right? And then I actually put the poem and said, this is Midjourney. It's reinterpretation is not just at the level of reinterpreting the image and how that image like I want to create like a Picasso.

I don't, I don't want my work to look like Picasso at all or anybody. I want my work to look like the Cubist movement mixed with the Fauvists mixed with the collages mixed with this, with … I want a new image to pop up. I want to see something brand new and that requires a lot of prompting, a lot of image prompting sometimes, a lot of different techniques.

And it's a trial and error kind of thing until you kind of find your way through. But that's a creative process. That's not hitting a button. That's not cutting and pasting or saying make this look like Picasso. That's something totally different.

Let’s take a moment to say thank you to our sponsor. “How to Fix the Internet” is supported by The Alfred P. Sloan Foundation’s Program in Public Understanding of Science and Technology. Enriching people’s lives through a keener appreciation of our increasingly technological world and portraying the complex humanity of scientists, engineers, and mathematicians.

And now back to our conversation with Nettrice Gaskins.

The way Nettrice talks about her artistic process using generative AI makes me think of that old cliche about abstract art – you know, how people say 'my kid could paint that.' There's a misconception now with Gen AI that people assume you just pop in a few words and boom, you get a piece of art. Sometimes that’s true, but Nettrice's approach goes far beyond a simple prompt.

Well, I did a talk recently, and it may have been for the Philadelphia Museum of Art. I did a lecture and the Q& A, they said, could you just demo? What you do, you have some time. And I remember after I demoed, they said, Oh, that definitely isn't hitting a button. That is much more, now I feel like I should go in there.

And a lot of times people come away, They're feeling like, now I really want to get in there, And see what I can do. Cause it isn't. I was showing, you know, in what, 30 seconds to a minute, basically how I generate images, which is very different than, you know, what they might think. And that was just within Midjourney. Another reason why personally that I got into on the prompt side before it was image style transfer, it was deep style. It wasn't prompt based. So it was about applying a style to. an image. Now you can apply many styles to one image. But then it was like, apply a style to this photo. And I spent most of my time in generative AI doing that until 2021, with DALL-E and Midjourney.

So before that, there were no prompts, it was just images. But then a lot came from that. The Smithsonian show came from that earlier work. It was like right on the edge of DALL-E and all that stuff coming. But I feel like, you know, my approach even then was somehow I didn't see images that reflected me or reflected, um, the type of images I wanted to see.

So that really propelled me into going into generative AI from the image style, applying styles to, for example, there's something if you're in a computer graphics major or you do computer graphics development or CGI, you may know a lot of people would know something called subsurface scattering.
And subsurface scattering is an effect people apply to skin. It's kind of like a milk, it's called glow. It's very well known, you texture and model your, your person based on that. However, it dulls dark skin tones. And if you look at photography and all the years with film and all that stuff, we have all these examples of where things were calibrated a certain way, not quite for darker skin tones. Here we are again, this time with, but there's something called specular reflection or shine, but apparently when applied, it brings up and enhances darker skin tones. So I wondered if I could apply, using neural image style transfer or deep style, if I could apply that shine or subsurface scattering to my photographs and create portraits of darker skin tones that enhanced features.

Well that succeeded. It worked. And I was just using 18th century tapestries that had metallics in them. So they have gold or they, you know, they had that shine in it as the style applied.


So one of those, I did a bunch of series of portraits called the gilded series. And around the time I was working on that and exploring that, um, Greg Tate, the cultural critic and writer, Greg Tate, passed away in 2021 and, um, I did a portrait. I applied my tapestry, the style, and it was a selfie he had taken of himself. So it wasn't like it was from a magazine or anything like that. And then I put it on social media and immediately his family and friends reached out.
So now it's a 25 foot mural in Brooklyn.


It's beautiful. I was looking at it earlier. We'll link to it.

Yeah, I’ve seen it too.

And that was not prompt based, that's just applying some ideas around specular reflection and it says from the Gilded Series on the placard. But that is generative AI. And that is remixing. Some of that is in Photoshop, and I Photoshopped, and some of that is three different outputs from the generator that were put together and combined in Photoshop to make that image.

And when it's nighttime, because it has metallics in there, there's a little bit of a shine to the images. When I see people tag me, if they're driving by in the car, you see that glow. I mean, you see that shine, and it, it does apply. And that came from this experimenting with an idea using generative AI.

So, and when people are thinking about AI right now, you know, we've really worked hard and EFF has been part of this, but others as well, is to put the threat of bias and bias kind of as something we also have to talk about because it's definitely been historically a problem with, uh, AI and machine learning systems, including not recognizing black skin.

And I'm wondering as somebody who's playing with this a lot, how do you think about the role bias plays and how to combat it. And I think your stories kind of do some of this too, but I'd love to hear how you think about combating bias. And I have a follow up question too, but I want to start with that.

Yeah, some of the presentations I've done, I did a Power of Difference for Bloomberg, was talking to the black community about generative AI. There was a paper I read a month or two ago, um, they did a study for all the main popular AI generators, like Stable Diffusion, Midjourney, DALL-E, maybe another, and they did an experiment to show bias, to show why this is important, and one of the, the prompt was portrait, a portrait of a lawyer. And they did it in all, and it was all men...

I was going to say it didn't look like me either. I bet.

I think it was DALL-E was more diverse. So all men, but it was like a black guy. It was like, you know, they were all, and then there was like a racially ambiguous guy. And, um, was it Midjourney, um, for Deep Dream Generator, it was just a black guy with a striped shirt.

But for Portrait of a Felon. Um, Midjourney had kind of a diverse, still all men, but for kind of more diverse, racially ambiguous men. But DALL-E produced three apes and a black man. And so my comment to the audience or to listeners is, we know that there's history in Jim Crow and before that about linking black men, black people to apes. Somehow that's in the, that was the only thing in the prompt portrait of a felon and there are three apes and a black man. How do apes play into "felon?" The connection isn't "felon," the connection is the black man, and then to the apes. That's sitting somewhere and it easily popped up.

And there’s been scary stuff that I've seen in Midjourney, for example. And I'm trying to do a blues musician and it gives me an ape with a guitar. So it's still, you know, and I said, so there's that, and it's still all men, right?

So then because I have a certain particular knowledge, I do know of a lawyer who was Constance Baker Motley. So I did a portrait of Constance Baker Motley, but you would have to know that. If I'm a student or someone, I don't know any lawyers and I do portrait of a lawyer for an assignment or portrait of whatever, who knows what might pop up and then how do I process that?

We see bias all the time. I could, because of who I am, and I know history, I know why the black man and the apes or animals popped up for "felon," but it still happened, and we still have this reality. And so to offset that one of the things is, has it needed, in order to offset some of that is artists or user intervention.
So we intervene by changing the image. Thumbs up, thumbs down. Or we can, in the prediction, say, this is wrong. This is not the right information. And eventually it trains the model not to do that. Or we can create a Constance Baker Motley, you know, of our own to offset that, but we would have to have that knowledge first.

And a lot of people don't have that knowledge first. I can think of a lawyer off the top, you know, that's a black woman that, you know, is different from what I got from the AI generators. But if that intervention right now is key, and then we gotta have more people who are looking at the data, who are looking at the data sources, and are also training the model, and more ways for people from diverse groups to train the model, or help train the model, so we get better results.

And that hasn't, that usually doesn't happen. These happen easily. And so that's kind of my answer to that.

One of the stories that I've heard you tell is about the, working with these dancers in Trinidad and training up a model of the Caribbean dancers. And I'm wondering if one of the ways you think about addressing bias is, I guess, same with your lawyer story, is like sticking other things into the model to try to give it a broader frame than it might otherwise have, or in the training data.

But I'm, I'm wondering if that's something you do a lot of, and, and I, I might ask you to tell that story about the dancers, because I thought it was cool.

That was the Mozilla Foundation sponsored project for many different artists and technologists to interrogate AI - Generative AI specifically, but AI in general. And so we did choose, 'cause two of my theme, it was a team of three women, me and two other women. One's a dancer, one's an architect, but we, those two women are from the Caribbean.

And so because during the lockdown there was no festival, there was no carnival, a lot of people, across those cultures were doing it on Zoom. So we're having Zoom parties. So we just had Zoom parties with the data we were collecting. We were explaining generative AI and what we were doing, how it worked to the Caribbean community.


And then we would put the music on and dance, so we were getting footage from the people who are participating. And then using PoseNet and machine learning to produce an app that allows you to dance with yourself, mini dancer, or to dance with shapes and, or create color painting with movement that was colors with colors from Carnival.

And one of the members, Vernelle Noel, she was using GAN, Generative Adversarial Networks to produce costuming, um, that you might see, but in really futuristic ways, using GAN technology. So different ways we could do that. We explored that with the project.

One of the things that, again, I'm kind of feeding you stuff back from yourself because I found it really interesting as you're talking about, like, using these tools in a liberatory way for liberation, as opposed to surveillance and control. And I wondered if you have some thoughts about how best to do that, like what are the kinds of things you look for in a project to try to see whether it's really based in liberation or based in kind of surveillance and monitoring and control, because that's been a long time issue, especially for people from majority countries.

You know, we were very careful with the data from the Carnival project. We said after a particular set period of time, we would get rid of the data. We were only using it for this project for a certain period of time, and we have, you know, signed, everyone signed off on that, including the participants.
Kind of like IRB if you're an academic, and in some cases, and one, Vernelle, was an academic. So it was done through her university. So there was IRB involved, but, um, I think it was just an art. Uh, but we want to be careful with data. Like we wanted people to know we're going to collect this and then we're going to get rid of it once we, you know, do what we need to do.

And I think that's part of it, but also, you know, people have been doing stuff with surveillance technology for a good minute. Um, artists have been doing, um, statements using surveillance technology. Um, people have been making music. There's a lot of rap music and songs about surveillance. Being watched and you know, I did a in Second Life, I did a wall of eyes that follow you everywhere you go...


NETTRICE GASKINS curate the feeling of always being watched. And for people who don't know what that's like it created that feeling in them as avatars they were like why am I being watched and I'm like this is you at a, if you're black at a grocery store, if you go to Neiman Marcus, you know go to like a fancy department store. This might be what you feel like. I'm trying to simulate that in virtual 3D was a goal.

I'm not so much trying to simulate. I'm trying to, here's another experience. There are people who really get behind the idea that you're taking from other people's work. And that that is the danger. And some people are doing that. I don't want to say that that's not the case. There are people out there who don't have a visual vocabulary, but want to get in here. And they'll use another person's artwork or their name to play around with tools. They don't have an arts background. And so they are going to do that.

And then there are people like me who want to push the boundaries. And want to see what happens when you mix different tools and do different things. And they never, those people who say that you're taking other people's work, I say opt out. Do that. I still continue because a lot of the work that, there's been so lack of representation from artists like me in the spaces, even if you opt out, it doesn't change my process at all.

And that says a lot about gatekeepers, equity, you know, representation and galleries and museums and all that thing are in certain circles for digital artists like Deviant, you know, it just, it doesn't get at some of the real gray areas around this stuff.

I think there's something here about people learning as well, where, you know, young musicians start off and they want to play like Beethoven, right? But at some point you find your own, you need to find your own voice. And that, that, that to me is the, you know, obviously there are people who are just cheaters who are trying to pass themselves off as somebody else and that matters and that's important.

But there's also just this period of, I think, artistic growth, where you kind of start out trying to emulate somebody who you admire, and then through that process, you kind of figure out your own voice, which isn't going to be just the same.

And, you know, there was some backlash over a cover that I had done for a book. And then they went, when the publisher came back, they said, where are your sources? It was a 1949 photograph of my mother and her friends. It has no watermark. So we don't know who took the photo. And obviously, from 1949, it's almost in the public domain, it's like, right on the edge.

So close!

But none of those people live anymore. My mom passed in 2018. So I use that as a source. My mom, a picture of my mom from a photo album. Or something from, if it's a client, they pay for licensing of particular stock photos. In one case, I used three stock photos because we couldn't find a stock photo that represented the character of the book.

So I had to do like a Frankenstein of three to create that character. That's a collage. And then that was uploaded to the generator, after that, to go further.
So yeah, I think that, you know, when we get into the backlash, a lot of people think, this is all you're doing. And then when I open up the window and say, or open up the door and say, look at what I'm doing - Oh, that's not what she was doing at all!

That's because people don't have the education and they're hearing about it in certain circles, but they're not realizing that this is another creative process that's new and it's entering our world that people can reject or not.

Like, people will say digital photography is going to take our jobs. Really, the best photography comes from being in a darkroom. And going through the process with the enlarger and the chemicals. That's the true photography. Not what you do in these digital cameras and all that stuff and using software, that's not real photography. Same kind of idea but here we are talking about something else. But very, very similar reaction.

Yeah, I think people tend to want to cling to the thing that they're familiar with as the real thing, and a little slow sometimes to recognize what's going on. And what I really appreciate about your approach is you're really using this like a tool. It's a complicated process to get a really cool new paintbrush that people can create new things with.

And I want to make sure that we're not throwing out the babies with the bathwater as we're thinking about this. And I also think that, you know, my hope and my dream is that in our, in our better technological future, you know, these tools will be far more evenly distributed than say some of the earlier tools, right?
And you know, Second Life and, and things like that, you know, were fairly limited by who could have the financial ability to actually have access. But we have broadened that aperture a lot, not as far as it needs to go now. And so, you know, part of my dream for a better tech future is that these tools are not locked away and only people who have certain access and certain credentials get the ability to use them.

But really, we broaden them out. That, that points towards more open models, open foundational models, as well as, um, kind of a broader range of people being able to play with them because I think that's where the cool stuff's gonna probably come from. That's where the cool stuff has always come from, right?

It hasn't come from the mainstream corporate business model for art. It's come from all the little nooks and crannies where the light comes in.

Yeah. Absolutely.

Oh Nettrice, thank you so much for sharing your vision and your enthusiasm with us. This has just been an amazing conversation.

Thanks for having me.

What an incredible conversation to have, in part because, you know, we got to talk to an actual artist about their process and learn that, well, I learned that I know nothing about how to use generative AI and that some people are really, really talented and it comes from that kind of experience, and being able to really build something, and not just write a sentence and see what happens, but have an intention and a, a dedicated process to making art.

And I think it's going to be really helpful for more people to see the kind of art that Nettrice makes and hear some of that description of how she does it.

Yeah. I think so too. And I think the thing that just shines clear is that you can have all the tools, but you need the artist. And if you don't have the artist with their knowledge and their eye and their vision, then you're not really creating art with this. You may be creating something, something you could use, but you know, there's just no replacing the artist, even with the fanciest of tools.

I keep coming back to the term that, uh, was applied to me often when I was younger, which was “script kitty,” because I never learned how to program, but I was very good at finding some code and using it. And I think that a lot of people think that's the only thing that generative AI lets you do.

And it's clear that if you have the talent and the, and the resources and the experience, you can do way more. And that's what Nettrice can show people. I hope more people come away from this conversation thinking like, I have to jump onto this now because I'm really excited to do exactly the kinds of things that she's doing.

Yeah, you know, she made a piece of generative art every day for a year, right? I mean, first of all, she comes from an art background, but then, you know, you've got to really dive in, and I think that cool things can come out of it.

The other thing I really liked was her recognition that so much of our, our culture and our society and the things that we love about our world comes from, you know, people on the margins making do and making art with what they have.

And I love the image of gumbo as a thing that comes out of cultures that don't have access to the finest cuts of meat and seafood and instead build something else, and she paired that with an image of Sashiko stitching in Japan, which came out of people trying to think about how to make their clothes last longer and make them stronger. And this gorgeous art form came out of it.

And how we can think of today's tools, whether they're AI or, or others as another medium in which we can begin to make things a beauty or things that are useful out of, you know, maybe the dribs of drabs of something that was built for a corporate purpose.

That's exactly right. And I also loved that. And I think we've discussed this before at EFF many times, but the comparison of the sort of generative AI tools to hip hop and to other forms of remix art, which I think probably a lot of people have made that connection, but I think it's, it's worth saying it again and again, because it is, it is such a, a sort of clear through line into those kinds of techniques and those kinds of art forms.

Yeah. And I think that, you know, from EFF's policy perspective, you know, one of the reasons that we stand up for fair use and think that it's so important is the recognition that arts like collage and like using generative AI, you know, they're not going to thrive if, if our model of how we control or monetize them is based on charging for every single little piece.

That's going to limit, just as it limited in hip hop, it's going to limit what kind of art we can get. And so that doesn't mean that we just shrug our shoulders and don't, you know, and say, forget it, artists, you're never going to be paid again.

I guess we’re just never going to have hip hop or

Or the other side, which is we need to find a way, you know, we, we, there are lots of ways in which we compensate people for creation that aren't tied to individual control of individual artifacts. And, and I think in this age of AI, but in previous images as well, like the failure for us to look to those things and to embrace them, has real impacts for our culture and society.

Thanks for joining us for this episode of How to Fix the Internet.

If you have feedback or suggestions, we'd love to hear from you. Visit EFF. org slash podcast and click on listener feedback. While you're there, you can become a member, donate, maybe pick up some merch and just see what's happening in digital rights this week and every week.

This podcast is licensed Creative Commons Attribution 4. 0 International and includes music licensed Creative Commons Unported by their creators.

In this episode, you heard Xena's Kiss slash Madea's Kiss by MWIC and Lost Track by Airtone featuring MWIC. You can find links to their music in our episode notes or on our website at slash podcast.

Our theme music is by Nat Keefe of BeatMower with Reed Mathis

How to Fix the Internet is supported by the Alfred P. Sloan Foundation's program in public understanding of science and technology.

We’ll see you next time.

I’m Jason Kelley…

And I’m Cindy Cohn.