Image: A Place for Tomorrow
Julie Kirsch via DALLE-2
In my last post, I considered whether AI art generators, like DALLE-2, Midjourney, and Stable Diffusion, pose an existential threat to artists and other creative professionals. I argued that they do, and my treatment of new AI art-making technology was largely negative. But as is the case with most new forms of technology, its consequences are not all bad. In this post, I shift gears and look at positive ways in which artists are embracing this technology and incorporating it into their creative practice. Drawing from the extended mind theory in the philosophy of mind, I also explore the possibility that human beings may be merging with artificial intelligence and other forms of technology.
As someone who already enjoys making pictures, I find DALLE-2, the art generator that I have experimented with, to be a lot of fun to use. It has an entertaining, game-like quality and involves an element of the unexpected. While you can influence the imagery that the generator produces by way of text prompts, you cannot control it in the way that you can an oil painting or graphite drawing. Sometimes you are disappointed by what it produces, but at other times you are delighted and inspired by its “inventiveness.” The challenge is to tame what is not completely under your control by mastering the language of text prompts. Through its use, which is now widespread, this technology has brought joy to countless people who would not otherwise be in the image-making business. It’s not difficult to appreciate the technology’s appeal. As Kevin Kelly explains, people are attracted to it simply because they like to look at pretty pictures:
“Like flames in a campfire, the light patterns are mesmerizing. They never repeat themselves; they surprise, again and again. They depict scenes no one has witnessed before or can even imagine, and they are expertly composed. It’s a similar pleasure to exploring a video game world, or paging through an art book. There is a real beauty to their creativity, and we stare much in the way we might appreciate a great art show or museum. In fact, viewing a parade of generated images is very much like visiting a personal museum—but in this case, the walls are full of art we ask for (“Picture Limitless Creativity at Your Fingertips”).”
In this way, the technology has ushered in a kind of democratization of image-making. People who would be too intimidated, insecure, or just plain lazy to pick up a pencil or paintbrush can use DALLE-2 and other AI art generators to produce images that are unique and responsive to their feelings, experiences, and preference.
Many professional artists have embraced developments in artificial intelligence in various forms as well. Artist Refik Anadol, for example, has integrated machine learning and data visualization methods into his technically and aesthetically impressive data paintings. His piece, Coral, is the product of over 100 million coral images of raw data. The artwork is the visualization of this data on a 40 by 40-inch screen—various forms of coral continuously ebb and flow into others (PBS). He created this piece, which was displayed at the World Economic Forum 2023, for the purpose of raising awareness about climate change. Unlike the casual DALLE-2 user, Anadol did not simply plug in a few descriptive text prompts into an art generator. Instead, he and the team of experts working in his studio (including computer and data scientists, engineers, researchers, and designers) collected the data themselves and translated it into the final piece using by programs that that they created and trained. Compiling the dataset alone was a years-long project (PBS/Refki Anadol). It is also worth noting Anadol and his team only use “ethically sourced” data that they themselves collect; that is, they do not “scrape” or steal images from other artists on the Internet when their compiling datasets (PBS). As Anadol sees it, his data paintings involve a collaboration between the human artist and the machine. In his view, the integration of AI into the artworld as a “beautiful movement” that brings art, science, and technology together; it is something that we should embrace since we are “getting closer to machines every single day.”
Anadol is not alone in sourcing his own images for his projects. Artist Sougwen Chung, who is also interested in human-machine interactions, designed and built her own AI robots to serve as art-making collaborators. Her early robots, all “Dougs” for “Drawing Operations Unit, Generation (followed by a number)” (Kaufman), were all trained on about 20 years of her own drawings. A more recent Doug, Doug 4, is connected to her via brain-wave data. Chung’s robots do not simply copy her drawings but offer new interpretations of her style in response to the dataset that they have been fed. In collaborating on a new painting, the robot responds to Chung, and Chung to the robot. They often do this in front of a live audience as performance art, the finished works selling for more than $131,000 (Kaufman).
Chung’s upbringing involved a synthesis of art and technology; her mother was a computer programmer and her father an opera singer. She played the violin and piano as a child and started coding and building websites in grade school (Kaufman). Chung is interested in exploring the interactions between smart technology and human beings. She does not subscribe to the various dystopian AI narratives that we often hear and is always looking for new ways to push her art and technology projects forward (Kaufman). She welcomes the technology and believes that interactions with AI devices, such as robots, provide us with new possibilities for creativity (TED).
Like Anadol and Chung, some high-profile artists embrace artificial intelligence in a way that goes beyond the causal use of mainstream AI generators like DALLE-2. Indeed, they are researchers as much as they are artists; often working with teams of experts, their technological accomplishments are as impressive as their artistic ones. But less tech savvy artists are also finding ways to integrate the new technology into their practice. For some artists, AI art generators, like DALLE-2, play a role in brainstorming or working through an idea. If an artist is not sure how to approach a piece, they can play around with it in an AI art generator. If an artist is interested in painting, say, a train crossing over a foggy river, they can plug these words into DALLE-2 (as text prompts) and see what it produces. In this way, the generator provides artists with materials to work with – “ideas” to explore in the creative process. The images that it produces may play the role that sketches or reference photos traditionally do in refining an idea or preparing for a new piece.
More dramatically, some theorists may take artists to be merging with the new AI technology that they have adopted, particularly if the artists in question consistently use a generator in their creative process. In the philosophy of mind, the extended mind thesis claims that the mind can extend beyond the skull to include external objects, such as smart phones and notepads, which play a regular role in our mental processing. In this view, associated with the work of Andy Clark and David Chalmers, a person’s iPhone becomes part of their mind when they offload memories, and thereby cognitive tasks, to it; in such cases, the iPhone plays a role in their mental processing that used to be played by the brain. While I do not take this view literally, I think that it provides us with a helpful way of thinking about the contribution that technology makes to our mental lives. Artists are “merging” with technology insofar as they are offloading imaginative tasks to AI art generators.
To appreciate the force of this point, compare how an AI generated image is produced with how a traditional artist’s sketch is produced. An artist who is formulating an idea for a painting may create a series of preliminary sketches. In these sketches, the artist may play around with value, texture, composition, and so forth. But importantly, the artist is the creative agent who produces and conceives of these sketches; the artist dreams them up in their imagination. In comparison, when an artist uses an AI art generator, the AI art generator is the one that “makes the magic happen,” as I put it an earlier post, and produces the imagery in response to text prompts. In this way, the AI art generator is stepping in for the human imagination and producing an image, or series of images, in response to text prompts.
If we give this a positive spin, then we may view this technology as providing artists with new possibilities for creativity, as Chung might put it. We can integrate machines into our creative process by offloading some of the imaginative work to them. We may even expand our imaginative capacities by doing this, for the machine may prompt us to take our initial idea in a new and unexpected direction. After reviewing the images that an AI art generator produces, an artist may revise their preliminary drawing and then set about working on the final piece. In this scenario, what the AI art generator produces is not the finished work, but new reference material from which to work. And given how common it is for artists to work from reference materials, this use of AI is nothing new or alarming.
It is possible that, in the future, technology like this will take the form of a neural implant, in which case the extended mind thesis will become even more compelling. Our own carbon-based brains will begin to mingle more closely with silicon-based implants. Neural implants already do exist to assist human beings in various ways. Incredibly, in 2022, Ujwal Chaudhary et al. reported that they had trained a man with CLIS (completely locked-in syndrome) to communicate via a neural implant simply by thinking. Their study showed that it is possible for someone who is completely paralyzed to communicate by using “brain-based volitional communication” (1). This technology will undoubtedly improve and expand in the years to come.
In my next AI post, I take up the issue of cheating and authenticity in art. We have already seen that AI art generators are beginning to play a role in the creative process of many artists. The use of this technology raises questions about what constitutes cheating in the art making process. In grappling with this issue, I will compare the way that an artist might use an AI art generator with the ways in which they use other tools and resources, such as reference images and Photoshop.
Works Cited:
Chaudhary, U., Vlachos, I., Zimmermann, J.B. et al. “Spelling Interface Using Intracortical Signals in a Completely Locked-in Patient Enabled via Auditory Neurofeedback training.” Nat Commun 13, 1236 (2022). https://doi.org/10.1038/s41467-022-28859-8
Chung, Sougwen. “Why I Draw with Robots.” TED. https://www.ted.com/talks/sougwen_chung_why_i_draw_with_robots?language=en. Accessed July 2, 2023.
Kaufman, Sarah L. “Artist Sougwen Chung Wanted Collaborators. So She Designed and Built Her Own AI Robots.” Washington Post. Nov. 5, 2020. https://www.washingtonpost.com/business/2020/11/05/ai-artificial-intelligence-art-sougwen-chung/. Accessed July 2, 2023.
Kelly, Kevin. “Picture Limitless Creativity at Your Fingertips.” Wired. https://www.wired.com/story/picture-limitless-creativity-ai-image-generators/. November 17, 2022. Accessed June 24, 2023.
“Use of Artificial Intelligence Generates Questions about the Future of Art,” PBS. https://www.pbs.org/newshour/show/use-of-artificial-intelligence-generates-questions-about-the-future-of-art. May 2, 2023. Accessed June 24, 2023.