The appreciation of works of art is subjective, and rightfully so as the experience of viewing art, and what it makes one feel is personal. This seems to guarantee that no artist or piece is left out of the realm of consumption, so long as there is a person who can appreciate the expressive work.
Nonetheless, one might still question about the qualifications of “art,” finding society’s attempts to define the elusive when sifting through the dictionary (or reading through the top hits on the internet) as guiding posts. Art has been deemed a skill acquired by experience, study, or observation, or instead the conscious use of skill and creative imagination especially in the production of aesthetic objects, or more relevantly, the expression or application of human creative skill and imagination, typically in a visual form such as a painting or sculpture, producing works to be greatly appreciated primarily for their beauty or emotional power. But, what happens to our understanding of art or our endeavors to craft definitions that capture its shifting nature if artificial intelligence (referenced throughout this piece in its colloquial name, A.I.) plays a more substantial role in its creation?
In 2022, Jason Allen won the first place prize for his piece titled “Théâtre D’opéra Spatial” in the “digitally manipulated photography” category of the Colorado State Fair, chosen over submissions from 21 other emerging artists. The piece that details three figures dressed in flowing gowns, staring out to either a bright beaming moon, or a window to the intensely lit beyond, depending on what your eye constructs, was created with the text to image A.I. tool, Midjourney.
Similar to its counterparts such as DALL-E, Midjourney takes “user-generated queries, runs them through an A.I. algorithm, and lets the algorithm pull from its source images and apply various artistic techniques to the resulting image.” However, unlike its predecessor DALL-E which has been touted as the go-to for more photo-realistic images, Midjourney is emerging as the preferred tool for more artistic creations, generating “images with complimentary colours, artistic use of light and shadow, sharp details, and composition with satisfying symmetry or perspective.” In effect, users such as Allen can dictate a prompt of “a woman in a Victorian frilly dress, wearing a space helmet,” sharpening his vision with the addition of words such as “opulent” and “lavish,” until the computer spits out an image that meets his expectations (along with those of judges in an annual state fair).
It should be noted that the intersection of art and technology is not necessarily new, and instead, creations arising from the surprising pair have existed for some time. For instance, in the 1960s, artists such as Jean Tinguely, inspired by cybernetic creations of the 1950s, exhibited painting machines that allowed visitors to choose the color, position of a pen, and the length of time to create an abstract work at London’s Institute of Contemporary Art. In 1973, artist Harold Cohen introduced “Aaron” to the world – a computer that he programmed to paint specific objects based on his own algorithms coded in his artistic style and which further exceeded his expectations by generating unimagined variations of the object. Further, as of today, generative adversarial networks (known as GANs), which use images as a sample set to deduce patterns that then aid in creating a new image, have become common tools in for artists seeking to engage with technology in their art.
While Allen’s piece can be grouped in this class of artistic A.I. creations, its placing top in the fair amongst submissions created solely by humans, under the presumption that this piece too was created by a human, has sparked conversations about what should be deemed art, and indirectly, what it means to create art. Perhaps this is because one of the noteworthy differences between Allen and artists who utilize A.I. as a tool of creation is the process. Rather than allowing a computer to calculate from its own database of images, other artists have themselves programmed the code used by a machine, or trained the algorithm with datasets of their own work, or fed it selected images from sites like Flickr (presumably with a license) that the computer then used to create the piece. It feels as if there is a more active control element demonstrated by these artists, as if their intent is more evident, whereas in the case of Allen’s creation, could this piece just be the product of chance?
The legal world itself appears to be grappling with the question of A.I. art as well. Just last year in February 2022, Stephen Phaler’s copyright application for the image titled “A Recent Entrance to Paradise,” which was created by his Creativity Machine, was rejected for the second time, failing to meet the threshold standard for copyrightability – i.e., being an original work of authorship fixed in a tangible medium. In its ruling, the United States Copyright Office (USCO) reasoned that “human authorship is a prerequisite to copyright protection,” and therefore, Thaler’s piece was ineligible for registration as Thaler did not assert that any human contributed to the process. Nor was Thaler’s argument that his Creativity Machine created the piece as a work-for-hire convincing on the reasoning that work-for-hire relationships arise out of legally binding contracts, which A.I. machines cannot enter into (at least for now), and the doctrine only determines who owns a piece rather than whether or not the piece is copyrightable.
Now, Allen’s case might have one notable distinction: he did offer prompts to the computer in order to create his first-place winning image, and maybe this might be enough to demonstrate human contribution. But even if that were the case, there’s still the issue that the database which these A.I.-generators are pulling objects, locations, people, and styles from are riddled with other artist’s works in the absence of that earlier mentioned involvement in the algorithm at the outset.  In effect, it stirs concerns around infringement, or at the very least, propels us deeper into the murky waters of the fair use doctrine. However, other countries seem to be expanding their own interpretations of authorship in the world of patents, as seen with Australia and South Africa, as a judge from the former ruled that A.I. creations could qualify for patent protection and the latter allowed Thaler to patent a creation autonomously invented by his A.I. machine. While the United States Patent and Trademark Office (USPTO) was less convinced by Thaler’s patent applications, it doesn’t rule out the possibility that the tide may turn further down the line with the increased engagement between A.I. and intellectual property.
While there’s still a ways to go before A.I. generators such as Midjourney will have any significant impact on the art world, there is something endearing about the USCO’s valuation of human touch – maybe a line should be drawn around our understanding of art if only to one day preserve our significance in a world where both humans and A.I. technologies co-exist. But putting our existential worries to the side, if art itself is theoretically evasive, should A.I.- generated art be left out of the definition? Let’s revisit those attempts to define art that were listed earlier:
Art is a skill acquired by experience, study, or observation. Allen shared in his Washington Post interview that he spent 80 hours curating 900 iterations of his piece, and the A.I. – generator he used is at its fourth version as of November 2022, increasing its knowledge of the objects that it features, its detail, and better compositions with just a single prompt.
Art is the conscious use of skill and creative imagination especially in the production of aesthetic objects. Allen prompted those 900 iterations with his own words, presumably arising from his own imagination, sharpening his prompts throughout the process, while the generator used its abilities to exact images until it fit his vision. Similarly, other artists who engage with A.I. to create art assist the precision of the algorithm as they “provide their input, guide the process, and filter the outputs” even going so far to utilize the algorithm to create a specific style.
Art is the expression or application of human creative skill and imagination, typically in a visual form such as a painting or sculpture, producing works to be greatly appreciated primarily for their beauty or emotional power. Allen did win the first-place prize for his image, so it would seem that his work evoked some emotional effect on those judges. Additionally, an A.I.- generated piece titled “Portrait de Edmond de Belamy” by the French collective Obvious, which was an experiment in whether “an algorithm [can] be creative,” marked a milestone in the timeline of A.I. – generated art when it was sold at Christie’s for $432,000 in 2018. There are also artists such as Mimi Onuoha who is further interrogating A.I.’s role in art with her series centered around “missing datasets,” highlighting the bias in algorithms, and in effect, generating libraries of these datasets that were once non-existent.
So, to be art, or not to be art? Well…it’s subjective.
 Art, Merriam-Webster Dictionary (n.d.), https://www.merriam-webster.com/dictionary/art.
 Art, Oxford Univ. Press (June 11, 2018), https://www.encyclopedia.com/literature-and-arts/art-and-architecture/art-general/art-history.
 Drew Harwell, He used AI art from Midjourney to win a fine-arts prize. Did he cheat?, Wash. Post: Tech. (September 2, 2022), https://www.washingtonpost.com/technology/2022/09/02/midjourney-artificial-intelligence-state-fair-colorado/.
 Mark Hachman, How to use Midjourney’s AI art generator, PCWorld (July 26, 2022), https://www.pcworld.com/article/820518/midjourneys-ai-art-goes-live-for-everyone.html.
 Guy Parsons, Everything you wanted to know about Midjourney, Dall•ery Gall•ery (July 25, 2022), https://dallery.gallery/prompt-resources-tools-ai-art/.
 Naomi Rea, How Did A.I. Art Evolve? Here’s a 5,000-Year Timeline of Artists Employing Artificial Intelligence, From the Ancient Inca to Modern-Day GANs, ArtNet (Dec. 16, 2021), https://www.encyclopedia.com/literature-and-arts/art-and-architecture/art-general/art-history.
 Ian Bogost, AI and Machine Learning Invade a New York Art Gallery, Atlantic (Mar. 6, 2019), https://www.theatlantic.com/technology/archive/2019/03/ai-created-art-invades-chelsea-gallery-scene/584134/.
 Rea, supra note 9.
 Jane Recker, U.S. Copyright Office Rules A.I. Art Can’t Be Copyrighted, Smithsonian Mag.: Smart News (Mar. 24, 2022), https://www.smithsonianmag.com/smart-news/us-copyright-office-rules-ai-art-cant-be-copyrighted-180979808/.
 U.S. Copyright Off., Second Request for Reconsideration for Refusal to Register A Recent Entrance to Paradise (2022).
 Harwell, supra note 4.
 Recker, supra note 13.
 Tim Schneider & Naomi Rea, Has Artificial Intelligence Given Us the Next Great Art Movement? Experts Say Slow Down, the ‘Field Is in Its Infancy’, ArtNet (Sept. 25, 2018), https://news.artnet.com/art-world/ai-art-comes-to-market-is-it-worth-the-hype-1352011.
 Harwell, supra note 4.
 Gowtham Raj, All You Need to Know About Midjourney V4 and It’s Use Cases, Decentralized Creator (Nov. 18. 2022), https://decentralizedcreator.com/midjourney-v4-and-its-use-cases/.
 Harwell, supra note 4.
 Rea, supra note 9.
 Naomi Rea, Is the Art Market Ready to Embrace Work Made by Artificial Intelligence? Christie’s Will Test the Waters This Fall, ArtNet (Aug. 20, 2018), https://news.artnet.com/market/artificial-intelligence-christies-1335170.
 Rea, supra note 9.