Story by Zachary Crain
Illustration by Emily Pack, using Shutterstock’s AI-powered image generator
UNC-Chapel Hill senior Yuqi Hu’s decision to pursue a career in website and graphic design was fueled by a simple ambition: she wanted to create.
“A lot of other fields, you have to follow the rules,” she said. “But art and technology, you have to be creative. I really like how we get to visualize things to engage people.”
Weeks from graduation, she’s spent her college career working to realize that ambition. For the last four semesters, she’s worked as a web designer and concept artist on the team at UNC Blue Sky Innovations — an organization in Chapel Hill that applies emerging technologies such as artificial intelligence to the media industry.
But as AI art generators — from OpenAI’s DALL-E 2 to Microsoft Bing’s Image Creator — have become more advanced, some in the graphic design community have been forced to consider how new technology could impact job opportunities and creative methods in the field.
Image generators work by converting written instructions from users into digital images using artificial intelligence.
“I just worry that AI is going to be something that can potentially take work away from working artists,” said Grayson Mendenhall, who teaches a course on media design at the UNC Hussman School of Journalism and Media.
‘They’ve gotten really good’
In March, German artist Boris Eldagsen was announced as the winner of the Sony World Photography Award. This month, he announced that he would not be accepting the award. The image was generated using AI.
Something similar happened last September when an AI generated image won a digital arts award at the Colorado State Fair Fine Arts Competition.
Circumstances like these — along with situations where people have sold AI-generated content — have been the source of ethical debates since the AI tools have become more available.
Mendenhall, who also works as a web developer and graphic designer at the UNC School of Education, said he doesn’t consider AI generated images to be art.
“In terms of ‘Did you take that photo, did you paint this, did you actually use a computer program to put those pixels in that exact same order on the screen,’” he said. “If somebody comes up with an image out of Midjourney or DALL-E or Stable Diffusion, I don’t consider that as a piece of art that they created.”
Ethical concerns aside, AI image generators are improving — and it’s happening rapidly.
Despite creative expression historically maintaining immunity to automation, it is becoming increasingly difficult to identify AI generated images.
“In the past year alone, they’ve gotten really good,” Mendenhall said. “I’ve got a feeling it’s going to keep growing and I don’t know where it really stops.”
Micah McLaughlin, a UNC junior journalism major studying graphic design, said he has quickly adjusted to the capabilities, and limits, of image generators.
“I do think that it can generate something in like 15 seconds that would take me hours to make,” he said. “But every time I do it, it’s not exactly what I was thinking of.”
Both Mendenhall and McLaughlin noted the lack of human control over an image’s outcome as a limit to AI’s abilities as the medium currently stands.
Though each of them said AI has its benefits in speed and as a tool to generate ideas, it lacks the detail that makes the best graphic designers.
“Every time you generate a prompt, you can’t exactly tell it to change minute details,” Mendenhall said. “(AI) can be a great starting point, but in terms of actually developing ideas, I think it helps if an artist can take it further themselves.”
Even with the current limitations, Mendenhall, Hu and McLaughlin all noted some concern about automation as AI platforms continue to develop.
“I try to avoid thinking about it, it’s kind of hard,” Hu said.
Steven King, founder and chief innovation officer at Blue Sky, said the potential for job loss needs to be monitored — despite his personal attempts at graphic design using AI generators not being as good as work from experienced artists.
“Throughout history, whenever new technology and new innovation comes, the way it’s pretty much sold is that you’re ultimately cutting the number of people doing that,” he said. “In the creative world that’s been avoided for a while, but that’s something we need to be concerned about and watch out for.”
‘Storytelling faster and easier’
Despite concerns over potential automation, King said AI can serve to create opportunities in digital storytelling.
Blue Sky is using AI in a project to build 3D avatars for virtual reality applications. But the potential uses of AI in design aren’t limited to high tech industries — they can have practical benefits on the day-to-day work of graphic designers.
One of the simplest uses of the new technology is helping artists find inspiration.
McLaughlin said that he has used AI image generators in the early stages of projects to test out different styles and find ideas he would have otherwise not thought of. Mendenhall said he has done the same, and that he sees benefits of using technology as a tool — even if it can’t create the detail necessary for a final product.
To King, improved efficiency is another major benefit that AI will introduce for artists and designers.
“I think there’s also some 3D generation that’s getting better, so I think that’s only going to grow in the near future,” he said. “Being able to see 3D models of exactly the type of character that you’re looking for that would’ve taken us days or weeks to build.”
King said AI technologies — beyond just image generators — can make “storytelling faster and easier.”
The increased efficiency, King said, still doesn’t take the need for creativity out of creative mediums.
“I think there’s a balance,” he said. “I think you’re going to see the need for creativity, but you’re also going to see prevalent use of AI that will replace a lot of jobs that are kind of that medium to low quality video and imagery that’s needed.”
‘I believe in human creative expression’
To hear Mendenhall tell it, fundamentals and creativity — the latter being same ideal that drove him generations of artists and creatives into graphic design — will continue to be a necessity to drive the medium forward.
“It’s always better to learn the fundamentals of something than to learn shortcuts,” he said. “Because if you learn fundamentals, you can always build things from the ground up.”
From the perspective of web design, Hu mirrored Mendenhall’s sentiment.
“AI needs stuff from humans to create,” she said. “I don’t think there will be a time where all the creative jobs will be gone because of AI, because they need us to be the base of that.”
King likened the relationship between AI and current forms of digital storytelling to that of camera phones and videography.
“You think about the number of people who now have access to a webcam or a camera phone, that’s totally changed the videography world and so a lot of people don’t think they need professional videographers,” he said. “But at the same time, we’re still hiring videographers to do all kinds of stuff that we need for creative work.”
Mendenhall has seen new technology disrupt an industry before. He started his career in the newspaper business. AI may have an impact on careers in other forms of storytelling next, but for Mendenhall, original art will always be a necessity. “I believe in human creative expression as something that’s positive in the world,” he said. “And AI is anything but human expression.”
