In the 1960s, artist Sol LeWitt came up with a new way of creating art. Instead of painting or sculpting, he started making “procedures” or “systems” that anyone could use to create a work of art. He called these procedures “prompts.” For example, one prompt might be: “Make a drawing using only straight lines.” Prompt engineering is now a key component in the creation of AI images. I wanted to see how an AI would interpret his prompts, so I fed them into image generators and was surprised by the results.
Sol LeWitt and his work as an engineer of prompts
Sol LeWitt was an influential artist unparalleled in his craft. Innovative and futuristic, he is most notable for his contributions to conceptual art, a term that he coined himself. As part of this movement, LeWitt focused on the idea behind an artwork rather than the intricate details of its creation. He provided artists with “structures,” or detailed assignments they could follow in order to produce pieces of art which embodied and effectively communicated their concept. He sought to make art accessible and encouraging people to create and view art without being confined by traditional formats and expectations. Being an engineer of prompts, Sol LeWitt formed creative pathways that allowed others to come up with artistic solutions through his own frameworks.
AI image generators and how they can interpret prompts much like Sol LeWitt’s work
Artificial Intelligence (AI) image generators create visuals based on prompts given to the AI by humans. Much like Sol LeWitt’s systems of instruction, these prompts provide the AI with algorithmic sets of information that are used to generate images in a wide variety of mediums. These computer-generated images can capture scenes from nature or everyday life as well as abstract shapes and forms. Some AI-generated images are realistic representations of their prompt, while others explore color, line, and technique unlike anything a human artist might create. AI image generators have opened up new possibilities for artists and designers looking for inspiration and ideas outside of traditional art practices.
Showcase some of the images created by AI in response to LeWitt’s prompts
As part of my Creative Coding class, I have my students interpret a Sol Lewitt prompt. This involves understanding the instructions given by the artist and then creating an image or animation in response. I am doing the same thing with AI image generators like Mid Journey and DALL·E 2, assigning them prompts formulated by conceptual artist Sol LeWitt.
The images created by AI in response to the prompts formulated by Sol LeWitt had mix outputs depending on how the prompt was phrased and the platform used. DALL·E 2’s outputs leaned more scientific and geometric while Mid Journey’s were much more “artistic” and “stylistic”. See the selected outputs below for three Sol LeWitt prompts generated on both Mid Journey and DALL·E 2.
Sol LeWitt Prompt 1
WALL DRAWING 295: Six white outlines of geometric figures superimposed on a black wall. View original at Mass MoCA.
Sol LeWitt Prompt 2
WALL DRAWING 56: A square is divided horizontally and vertically into four equal parts, each with lines in four directions superimposed progressively. Line drawing on wall. View original at Mass MoCA.
Sol LeWitt Prompt 3
WALL DRAWING 260A: On blue walls, all two-part combinations of arcs from corners and sides, and straight, not straight, and broken lines within a 36-inch (90 cm) grid. View original at Mass MoCA.
The images created by AI in response to the prompts formulated by conceptual artist Sol LeWitt are fascinating. The results were stunning, with AI creating new pieces of art that straddle the line between traditional styles and modern digital abstraction. With the potential for art reaching far beyond the limits of reality, these mesmerizing images help us see how AI can lead us to something completely unique and completely original.
What do these images reveal about the relationship between human creativity and AI generated images?
The recent emergence of AI generated images has highlighted our fascination with art and technology. On the one hand, these images reveal how our creative processes as humans can be enhanced by technology — allowing us to explore limitless possibilities in a way that has traditionally been impossible through traditional methods. On the other hand, it demonstrates how we have become increasingly reliant on artificial intelligence for artistic output. In the end, it is clear that the relationship between human creativity and AI generated images will continue to evolve over time, introducing new and fascinating ways for both human and machine to explore their own unique pathways towards artistry.
In many ways, the images created by AI in response to Sol LeWitt’s prompts are reminiscent of his own work. Both are constructed from a set of rules or constraints, and both emphasize process over product. However, the images generated by AI also reveal something about the potential future of human creativity. As AI becomes more sophisticated, it will be able to increasingly interpret and respond to the same kinds of prompts that have long been used by artists as a way to generate new ideas. In doing so, AI may someday come to play a role in prompting and shaping human creativity itself.
View the work of Sol Lewitt at Mass MoCA’s Sol Lewitt: A Wall Drawing Retrospective.
This post was written in collaboration with Jasper.ai.