Critical AI

Critical AI – The Ocean Gives Birth

Janis Palma

Is it art?

     There is an inescapable question when we are confronted with computer-generated graphics: can art be created using artificial intelligence (AI)? If so, what sort of AI platforms are best suited for this human activity? If not, then what do we call the end-product of an AI platform prompted by a human’s creative imagination?

Scholars in the intersections of art and technology seem to agree that, historically, these two have not been distinct and separate activities at all. As Joanna Zylinska puts it, “for ancient Greeks, there was no distinction between art and technology.”[1] The origins of the word artificial as derived from artifice, “in its Latin etymology (artificium) . . . signal[s] art, craft and skill.” [Emphasis mine.] Art historian Laurence Bertrand Dorléac takes it even further back in time to the prehistoric cave paintings, pointing out how “artists have remained connected to their technical environment” by appropriating the tools immediately available to them.

     To the extent that we can use algorithms designed by scientists like we use fine art paint brushes designed by anonymous craftsmen/women as our tools “to invent new forms,”[2] then the final work of art is indeed the artist’s own and not attributable to the instrument used to create it. The brush obeys the artist’s hand movements as much as the AI platform obeys the artist’s prompts. The medium will do what the medium does, whether it’s clay, oils, marble, or a computer algorithm. Even with photography as a technological medium for the creative arts, the artist’s manipulation of the medium, as in the case of Diego Trujillo Pisanty’s sound-to-image camera,[3] is what ultimately results in a work of art. It follows, then, that if the human hand—or the human mind—does not master the media, the creative impulse will be unsuccessful and no work of art will result, no matter how many attempts the aspiring artist makes. This is just as true for a charcoal drawing, a metal sculpture, a digital photograph, as it is for an AI-produced image.

Is the medium the message?

     In 1964 Marshall McLuhan popularized the phrase, “the medium is the message,”[4] suggesting that the “medium through which a message [was] conveyed [became] more important than the content of the message itself.”[5] Ontologically, however, we cannot divest the medium from its human genesis. Furthermore, in her analysis of AI-produced art, Yuk Hui underscores the dangers of attributing a protagonist role to the novelty of AI while dismissing traditional art forms as obsolete, because “such a classification is only based on a superficial reflection on art as mere artifact.”[6]

     When I experimented with Runway, an AI platform whose “mission is to build expressive and controllable tools for artists that can open new avenues for creative expression,”[7] I wanted to create something that said, the ocean speaks to me. By critically engaging with this new medium, I envisioned a final piece that would “[say] something different about the world.”[8] My tools included a short video of the ocean I had recorded, and Runway’s Generative Video function. I typed in my prompt but, much to my disappointment, what I got back was nowhere near what I wanted this AI to produce. I tried new prompts, building up on what the platform had already gotten wrong, adding minor variations each time, hoping the repetition would trigger an algorithmic recursion. “In computer science, a recursive function is one that calls itself during its execution.”[9] and generates “a complexity that is beyond iteration (mere repetition), because it consists of various spiral loops instead of only one mechanistic, repetitive loop,”[10] which can also be consistent with Aristotle’s conception of art as “mimetic, but mimesis, proceeding by addition and not just repetition.”[11] However, neither hand nor mind possessed any mastery over the tools available to me, and so the Runway responses to my prompts were likewise lacking in artistic accomplishment.

     The next time I tried to work with Runway I inadvertently opened a different AI platform called aitubo, a “Runway AI video generator.”[12]  It was a serendipitous invitation to try a different tool and see what my prompt could do. This time I added more details and what aitubo produced was a visually beautiful yet “quite superficial”[13] variety of art with some mesmerizing movements and garish colors[14] but certainly not anything I could claim as my own vision of what I wanted to create. Whatever either of these AI platforms produced was not saying “the ocean speaks to me.” Nevertheless, I did put all those fragmented images together using Microsoft ClipChamp and created something I decided to call “The Ocean Gives Birth.” The name is a metaphor for what happened to the video I had recorded, not so much for the AI-produced images.   

Conclusion

     To paraphrase the catalogue for the 2018 exhibition of Artistes & Robots at the Grand Palais in Paris, “[AI does not] replace the artist or art: [it] invite[s] us to ask what makes a work of art – and what makes an artist.”[15] Even Jackson Pollok had a personal aesthetics that guided his apparently random splatters of paint. George Braque and Juan Gris—among many others—were very intentional in their fragmentation of two-dimensional imagery, not as a way of integrating technology but as a way of defeating its frontal attack on the requisite “humanist notions of originality and genius.”[16] I would therefore answer my initial question—can art be created using artificial intelligence?—with a very firm “yes,” but only as long as there is a talented artist inserted in the creative process from beginning to end.

Endnotes: 

[1] Zylinska, Joanna. (2020) AI Art: Machine Visions and Warped Dreams. London: Open Humanities Press.

[2] Zylinska, citing Dorléac, Laurence Bertrand and Jérôme Neutres. Eds. 2018. Artistes & Robots. Paris: Réunion des musées nationaux – Grand Palais.

[3] https://www.creativeapplications.net/sound/blind-camera-point-and-shoot-sound-to-image.

[4] McLuhan, Marshall. (1964) The Medium is the Message. New York: Bantam Books.

[5] ‘The Medium is the Message.’ August 17, 2023. Mcluhan.org. https://mcluhan.org/the-medium-is-the-message. Last visited on 28 November 2024.

[6] Hui, Yuk. (2021) Art and Cosmotechnics. Minneapolis: University of Minnesota Press.

[7] https://runwayml.com/research/introducing-act-one.

[8] Zylinska, 49.

[9] Hui, 233.

[10] Hui, 234.

[11] Zylinska, 50.

[12] https://app.aitubo.ai.

[13] Zylinska, 49.

[14] Ibid.was 

[15] Zylinska, 57. Citing Dorléac and Neutres 2018, 59.

[16] Zylinska, 50.

References:

Hui, Yuk. (2021) Art and Cosmotechnics. Minneapolis: University of Minnesota Press.

McLuhan, Marshall. (1964) The Medium is the Message. New York: Bantam Books.

‘The Medium is the Message.’ August 17, 2023. Mcluhan.org. https://mcluhan.org/the-medium-is-the-message. Last visited on 28 November 2024.

Zylinska, Joanna. (2020) AI Art: Machine Visions and Warped Dreams. London: Open Humanities Press.

Prompt: Wide angle shot of stormy dark ocean surface with 6-foot waves and white caps, POV moves underwater and zooms in to a fluorescent pink coral reef with multicolored fish swimming around it