Aurora Wilder

A Curatorial Statement

An Ambiguous Place: A Photographic Exploration of
Art and Artificial Intelligence’s Affiliation
— By Emily Edwards, Associate Curator, Dallas Contemporary

For centuries, artists have sought out new and unconventional art mediums and tools to express their artistic philosophies. Photographers Patrick Corrigan and Jennifer Pritchard have placed themselves within this lineage in art history by joining forces with an AI-image generator, DALL-E, for their An Ambiguous Place project. Employing this cutting-edge technology, they challenged AI to a creative faceoff to survey what images could be created when responding to abstract prompts and how it could expand their own personal practices.

In early fall of 2022, Patrick Corrigan and Jennifer Pritchard combined forces to form the artist collective known as Aurora Wilder. Intrigued by the evolving abilities of AI-image generators, they set out on a venture to work with an AI program to produce daily “photographs” inspired by prompts they drafted and then fed to it. They purposefully chose to work with DALL-E, a “generative AI technology that enables users to create new images with text to graphics prompts,” because of its ease of use and accessibility to all, with no need for users to possess programming skills. Developed by OpenAI in January of 2021, the deep learning model’s name was chosen for its pronunciation, similar to Dalí, as in the legendary Surrealist artist Salvador Dalí. DALL-E was trained on approximately 650 million image-text pairs found on the internet, and the collective aimed to work with the tool to dream up together new works pulled from this large pool of references. Daily, either Pritchard or Corrigan would feed DALL-E a prompt for its illustrative inspiration. Pritchard’s prompts would veer toward the existential while Corrigan’s would question reality, but both would retain a poetic yet vague voice that would give the system room to experiment. DALL-E would then produce a photograph in return. Simultaneously, the artists would go out in the world and take a photograph responding to that same prompt they had submitted. In the coming months, they would then pair together the two images, one human made and one AI-made, with the prompt and archive the day’s output, allowing for direct comparisons to be made between the photos. By the end of 2022, the project’s production was complete, and the innovative and bold undertaking would be titled An Ambiguous Place.

The images created by DALL-E and the photographs taken by Pritchard and Corrigan are interesting to compare. Both are mainly landscape and nature shots, non-figurative with few exceptions, and contain moody and atmospheric lighting. The collective instituted agreed-upon ideas around moods, stylistic cues, and keywords to guide the output created by both themselves and the generator. On October 10, 2022, the kick-off to the project, Pritchard wrote the prompt, “Reflections of the middle place between reality and dreams, paradox of the immense and the small.” She then took a photograph of what appears to be a window looking outdoors, with vivid green leaves of a bush or tree scaling up the image’s sides while the center shows a reflection of an interior setting featuring a chair, bookshelf, and a door. DALL-E produced an image that differed wildly from Pritchard’s, with the shot containing mostly an image of water rippling and reflecting a building that emerges in the top right corner, brightly lit with the shadow of the architecture taking over a third of the shot. The two images are vastly different in lighting, subject, and treatment, yet both match the cue given and are striking photographs. As time passed and the project progressed, the DALL-E and artist images grew closer in resemblance. On December 7th, Corrigan gave the prompt, “Keeping the waves as a constant.” The artist created and edited an image of a body of water under a celestial sky, with swirls of rainbow colors mixed in with the clouds above the low-hanging moon and reflected in the dark sea below. The horizon line is edited to look like the shot is almost bifurcated and rearranged. DALL-E produced a shot of the ocean at night, the water the color of midnight with the sky above a deep blue with spread out clouds above. Although not featured in the image, the water glows from unseen moonlight. Both shots are similar in subject, time of day, and coloring. Throughout the project, there are moments where one questions if both images were produced by the artists as DALL-E increased its skill level and gained a firmer grasp of the abstract prompts it was fed. The series of images express wonder, loneliness, ennui, curiosity, and respect for beauty. For an unfeeling AI generator, it’s remarkable how human the work becomes.

It is interesting to note that photography itself, the chosen medium of Corrigan and Pritchard, was once as controversial and questioned as AI-generated art is today. Since William Henry Fox Talbot declared in 1835 that the calotype of his estate was drawn by the house itself, photography was seen as synonymous with the eradication of human intervention since the machine is making the images technically. Over the next century, legal and creative battles were fought to establish photography as either a creative human expression or an objective mechanical documentation, with the general consensus landing somewhere in the middle. Artists worked with the new technology of photography as a tool to create more realistic images that could be mass produced, but their eye and direction formed the resulting photo. Alfred Stieglitz, the father of modern photography, put it best when he said, “After the selection of the subjects, the posing, lighting, exposure and development, every succeeding step…required little or no thought.”

Artists have also had assistance in the production process for generations. In the 1700 and 1800s, artist apprentices would mix pigments, prep and tape canvases, and even paint the backgrounds and subsidiary characters for their masters. Major contemporary artists such as Damien Hirst and Jeff Koons rely on a robust studio system of assistants to do the bulk of the actual sculpting to match the demand for their work. In the past century, artists have also incorporated industrial production to outpace their human capacities through photography, video, digital media, the internet, NFTs, and now AI. The idea of the “artist’s touch” has long been complicated by aides, both human and machine, deputized to fabricate on behalf of the main creative. These advancements liberate artists from pure factual representation and allow for them to focus on experimentation with color, light, movement, figuration, substance, context, and dimensionality, overall strengthening the output of work into the artistic field. By supplying DALL-E with an idea for an image via their prompts, Corrigan and Pritchard’s An Ambiguous Place project situates itself within this lineage in art history by taking the latest technology tool and using it as a creative collaborator.

After DALL-E produces an image, it prompts its user if its attempt was successful. In the weeks and months of working with the generator, DALL-E became more accustomed to Corrigan’s and Pritchard’s needs and wishes for the images. Both artists amended prompts occasionally to allow for DALL-E to get closer to the desired photograph. This push-and-pull between the artists and the technology they utilized for the project bettered both man and machine’s creative output. As the photographers refined their prompts to DALL-E, they honed in on the larger meaning of the overall series. AI generators help its viewers understand that all images contain multiple meanings that come from a culmination of data that is accrued during a lifetime of experiences. But DALL-E cannot actually make meaning through images itself; that’s where Corrigan and Pritchard step in and supply it to the work. Artists have always grappled with what an image is actually intended to do. By working with an artificial intelligent system, the collective pushes the definition of a work of art to its limits. An Ambiguous Place quite literally explores the middle zone of what art actually is: Is it the process or the human component or the discussion that emerges from it? This important project by Aurora Wilder collective allows for these questions to unfurl and gives room for their exploration.

Reach out to learn more.

^{1) Sean Michael Kerner, “What Is Dall-e (Dall-e 2) and How Does It Work?,” Enterprise AI, April 21, 2023, https://www.techtarget.com/searchenterpriseai/definition/Dall-E

2) Ibid.

3) Ramesh, Aditya, Prafulla Dhariwal, Alex Nichol, Casey Chu, and Mark Chen. “Hierarchical Text-Conditional Image Generation with Clip Latents.” arXiv.org, April 13, 2022. https://arxiv.org/abs/2204.06125.

4) Rosson, Lois. “What Is Ai Doing to Art?” NOEMA, April 11, 2023. https://www.noemamag.com/what-is-ai-doing-to-art/.

5) Ibid}