AI-powered Image Tools Create An Existential Crisis for Visual Artists
AI-powered image generators create new challenges for visual artists to surmount.
With AI-powered tools producing images that would be classified as art, existential crisis hangs like a shadow over visual artists. Brian Jackson, research director at Info-Tech Research Group, examines the creative potential of generative AI and what that means for the future of visual art and artists.
Visual artist Rob Sheridan once developed a narrative concept for the industrial rock band Nine Inch Nails depicting the end of the world under the boot of an authoritarian regime. Yet it wasn’t until he saw the creative potential of artificial intelligence-powered image generators that he had an existential crisis.
Sheridan had early beta access to Midjourney, an image generator tool currently only accessible through the Discord chat application. The self-described research lab behind the tool is helmed by a former NASA researcher and advised by a handful of Technorati elite out of Silicon Valley. Sheridan found that with the mere typing of a text prompt, the AI tool generated beautiful and illustrative images that could then be iterated upon and refined. It was as if the concepts he could only previously imagine in his own mind were now externalized and brought to life, albeit in a distorted and imperfect form.
“It made me question my role as an artist,” he says in an interview. “How can I have authorship in this process when the rendering is done by the computer?”
The crisis was brief. After a weekend of honing his “prompt craft” to get the images he wanted, Sheridan became excited about the possibilities. He plunged into the subject matter where he was most comfortable – horror-genre nightmares. He also noticed that when asked to produce images in the style of old photographs, the quality took on a sort of uncanny valley effect. Sheridan became drawn to the “wrongness” of the AI.
“I love the flawed, messed up aesthetic of it,” he says of Midjourney’s algorithm. “It’s great if you want to make a ******-up tentacle keyboard.”
Out of that combination, Sheridan started putting together a graphic narrative. He created Instagram and Twitter accounts named the Volstof Institute for Interdimensional Research. He wove together AI-generated images with notes from fictional scientists, a combination of experimental observations, and a personal diary. The story of a secret research laboratory that studied the forbidden liminal space between two dimensions was told with a “found footage” sort of feel.
WARNING: What lies ahead is a nightmare from the darkest shadows of scientific history that will shatter your understanding of the universe. And we've only just begun.
Welcome to VIIR. All of our findings so far are catalogued here, in order of discovery: https://t.co/UEhH6Tdsyj
— Volstof Institute for Interdimensional Research (@VolstofResearch) April 2, 2022
In describing his sci-fi horror genre graphic story published on Instagram, Sheridan compares its aesthetic to that of M.C. Escher and H.R. Giger. While he didn’t have the opportunity to collaborate with the late artists, his knowledge of their artistic influence allowed him to tap into it through Midjourney’s artificial intelligence.
See More: How Google’s Sunset of Conversational Services Impacts the Way I Change Diapers
Generative AI Gives Rise to a Powerful New Creative Capability
The artificial intelligence behind Midjourney and other image generators coming to the market over the last several months is in the generative AI category. It works by using a neural network trained to produce content based on a massive amount of data. An adversarial neural network is used to discriminate between good quality output and garbage. The resulting content is meaningful to people, but the implications of generative AI will affect many different creative fields in time.
Sheridan’s work on the Volstof Institute for Interdimensional Research was done with the Midjourney bot. Despite only being available through a Discord server, the service’s popularity is taking off like a rocket ship. At a mid-August office hours session held on Discord, moderator DavidH spoke about the problem of coming up against the limits of a Discord server.
“If it wasn’t for the million-person server limit, we’d be bigger than the Minecraft and Fortnite channels combined,” he says, referencing chat rooms for two popular video games. “I don’t think that’s happened before.”
To pare back the channel size to fewer than 1 million users, moderators kick people out of the channel daily. While spammers are a common nuisance on Discord and an easy choice to cut, the threshold for who gets to stay in the community is much higher.
Midjourney is just one AI-image generating service available. Others, like Craiyon.com, are more easily accessible via a web browser. In July, the creator of that website told CNN that users were typing in 5 million prompts per day to create images. Accomplice.ai is a startup founded in September 2021 that’s raising venture capital to build a business model that would compete with stock photos and art websites, providing access to a variety of different models in one integrated interface. Stable Diffusion is a new model that became available to the site in mid-August. Having just exited beta, the model is catching attention for being open source and offering good processing speed and image quality. Two weeks after it was available to Accomplice.ai, users on that site had already created 10,000 images. Both Midjourney and Accomplice.ai use a freemium business model, offering users a limited number of images generated for free but unlocking full access for a monthly fee (US$10 in Midjourney’s case).
The images generated by these services are good for much more than a quick hit on social media. In August, Jason Allen won Colorado State Fair’s fine art competition with a piece generated using Midjourney, titled Théâtre D’opéra Spatial. Some artists claimed Allen was deceptive. On Twitter, some users proclaimed the death of artistry. Despite the popularity of the services, many artists had concerns that it be represented in a way that is equal to human works.
Jason Allen's AI-generated work "Théâtre D'opéra Spatial" took first place in the digital category at the Colorado State Fair. https://t.co/6bFNFERCki
— VICE (@VICE) September 2, 2022
The backlash against the services has grown in the last few weeks. Online artist communities have banned images created by the AI bots after being flooded with the works. Newgrounds, Fur Affinity, and InkBlot have all declared bans on AI art with limited exceptions.
The debate around AI image generators is in full swing, not just in terms of the merit artists deserve when using them. Critics have pointed to bias in the algorithms, indicating ingrained sexism or racism. Others point out the ethical problem that image-generating algorithms were trained using millions of images scraped from the internet. The original owners of those images are not receiving credit or compensation for their work.
Still, many are embracing the new tools and considering how to put them to work. A user in the Midjourney open office hours shares that he plans to incorporate the tool into his visual media class. He sees it as a way to break down the barriers to creativity, allowing those who can’t paint or draw well to still participate in character design.
See More: Is Synthetic Data Set To Become a Mainstream AI Strategy?
A New Form of Literacy
Moderator DavidH says that Midjourney is spreading artistic awareness and literacy. Because the knowledge of influential artistic styles, mediums, and perspectives is helpful in writing prompts to produce the image intended by the user, it encourages learning about art history.
“It’s something different from what has come before,” he says. “There was a time in society when only some people could read and write, and now we have a society based on the assumption everyone can read and write.
“We’re at that point with visual expression. We’re moving from a society that can’t look at things or create things in a visually sophisticated way to one where they can.”
Stories of artists having an existential crisis with generative AI are filling the headlines today, but it’s just the first of many fields that the technology will disrupt. The same algorithmic process used to create new images can also be applied to composing music, writing research papers or news stories, and coding software.
When the creators in those areas find themselves wondering how to react to the disruption, they might want to heed Sheridan’s advice – adapt.
“In order to make something that stands out, it does take a lot of creativity and input from a skilled artist,” he says. “I viewed working with Midjourney’s bot as a collaboration with the AI. Let the boundaries and the parameters guide the process.”
Creators are used to collaborating with their peers. Now they’ll just have to learn how to do it with AI too.
Can visual artists collaborate with AI-power image generators to evolve together? Tell us what you think on Facebook, Twitter, and LinkedIn. We’d love to know your take on this!
MORE ON AI
- AI Adoption Will Grow Ten-fold by 2030, Dominated by Use in IoT
- New Research Identifies Key Ways that Enterprises Will Use AI for Competitive Advantage
- How Conversational AI Is Enabling Brands To Move beyond Third-party Cookies
- From Raw Data to ML Models: The Magic Behind AI-Powered Feature Engineering
About Expert Contributors: The Expert Contributor program is designed to help kickstart meaningful conversations around the priorities and challenges most critical to C-level executives. The insights and perspectives will help CIOs tackle what’s most important to them. We are always looking for industry thinkers who can help set the narrative for our enterprise audience. To know more about this program, and submit your ideas, reach out to the Spiceworks News & Insights Editorial team at editorial-toolbox@ziffdavis.com.
