AUSTIN (KXAN) — As artificial intelligence becomes more and more prevalent, one University of Texas at Austin lecturer is preparing his fine arts students for a changing art industry after graduation.
Tyler Coleman, an adjunct lecturer in UT’s College of Fine Arts and a gaming developer, shared with KXAN more about how he is teaching his students about AI use and the future of AI in art.
Coleman said he is part of a UT task force to determine how to best handle AI. He suggested other educators find ways to embrace AI in their own platforms.
“All of these schools are suddenly rushing to figure out how to respond to this, the same way they did to Wikipedia 15 years ago, right? The same way that they have to calculators in your phones, right?” Coleman said.
Editor’s note: This interview was edited for length and clarity.
Q: How are you seeing AI used in art?
Tyler Coleman: There’s been a long, long-standing art movement, usually referred to as algorithmic or generative art. It’s been around for decades—visual art, music, installation pieces. The recent advancements in machine learning, which now everybody’s calling AI, have led to a sort of explosion of creativity and exploration of that space.
It started with things like Deep Dream. Then, it was a style called GAN, which was in style transfer. So you could take a photo of yourself—take Vincent Van Gogh’s art—and make your photo look like it was made by Van Gogh.
In the image space, there’s been this massive explosion of these tools combined with the fact that now we see these large language models. The combination of being able to generatively create text and image is leading to people exploring creative spaces that they otherwise didn’t,
There are a few spaces that are still unexplored, but certainly will be within the next few years.
Q: What are the ethical differences between creating digital art and using AI to create art?
Coleman: The most immediate question is whether the data used to train these models needs to be copyrighted. Is there a fair use doctrine use of this? Whether the output is then considered copyrightable is a big question. So there’s a lot of legal questions there.
The ethics question, at least for the way I teach the class is very much a subjective question. Ethical guidelines tend to differ by person, culture and generation. So we more or less opened the question to the class. We have discussions on what ethics means to them, what they find ethical and unethical, and then we go through a whole series of projects where they create ethics projects related to AI. So they researched everything from deep fakes to how AI will impact education to AI and machine learning in the medical system.
So, we as a class picked a bunch of topics and explored those spaces. Because certainly, it’s going to impact these areas. Whether you feel that ethically you’re for or against it, someone will feel otherwise and probably explore the space.
Q: When did you start having these conversations with your students?
Coleman: In Spring 2022, I taught my first fully AI-focused course. I’ve sort of shown a little bit of what it could make all the way back to 2019 to 2020, and I’ve taught classes on the idea of using code to create art. But in the past year, certainly, it’s expanded.
My process tends to show people the tools, let them play around with them…understand its strengths and weaknesses, what it does well, and what it doesn’t.
If it’s going to be a whole new world when students graduate, the best that you can do is prepare them for that world.
Q: Are your students excited, concerned or questioning what the future of their career looks like?
Coleman: There’s a self-selection bias. They choose to be in my class. So the students that choose it, generally are cautiously optimistic.
There’s two or three students that I would say maybe you’re just full-on gung ho excited for it. For the most part, especially after the ethics section, they’re all like they understand the risks. But they also understand that there’s nothing that they, as current college students can do at a very large scale to prevent this from happening. So the best thing they can do is learn about it, prepare for it and be the people who are experts in these spaces.
I have seen rumblings of concern. I had a few students tell me ‘I refuse to take your class because I don’t want to learn about the thing that’s going to replace me.’
In my mind, it’s very much knowing your enemy and knowing what’s coming. If you sort of choose to stick your head in the sand or largely ignore it. It doesn’t stop it from happening.
Q: How do you see AI impacting the gaming and art industries moving forward?
Coleman: I think there’s going to be a lot of places that either embrace it or forbid it within their production cycles. Those who outlaw it will use that as part of their branding. It will be like a stamp that says ‘we used no AI in the creation of this,’ and that will be part of their marketing process. They will be at a disadvantage for not using some tools that exist. So they need to find that advantage elsewhere.
I think we’re gonna see that rift starting this year. We’re already seeing a little bit of rumblings. Some companies have announced ‘like ‘we’ll never allow that content on our platforms.’ Some are saying ‘we are allowing it, but we need it to be tagged.’
Q: How are you using AI in your work?
Coleman: We have a sort of internal mantra of using it for pre-production. We use it for just ideation, thinking of things, generating hundreds of images, lists of names, rigidity, and then sort of use it as one more brainstorming partner.
My sort of attitude is if we use it at the beginning, we’re not necessarily replacing jobs. It’s kind of benefiting those who are already in that space. A small studio like mine can’t afford people whose sole role is that beginning sort of brainstorming concept artists and prototype errors. So we use it to sort of increase our productivity as a small team.
If you were allowed to take a selfie in my game, and it creates a character based on you, it’s simply not possible to have thousands of artists quickly draw that and send it back to you when you do it. So looking at AI in places where it expands functionality that otherwise couldn’t be done.