Ai-Da Robot and the Future of AI “Artists”
Ai-Da Robot and the Future of AI “Artists”
In February 2019, “the world’s first ultra-realistic artist robot”, named Ai-Da, was created by Aidan Meller. She partially resembles a real person, including realistic skin, hair, and clothes, but her limbs are purely mechanical. The website about her, aidarobot.com, states that she is “a machine with a fusion of electronic/AI/human inputs” that uses cameras in her moving eyes, algorithms, and her robotic arm to draw and paint. Moving independently, Ai-Da creates compelling works of art, one of which was recently sold at auction for $1 million. Although she represents a feat of technology, the implications and future possibilities of AI art since Ai-Da’s debut provoke mixed feelings and an air of uncertainty especially for human artists.
Are machines like Ai-Da more ethical than AI image generator programs? It is unclear what data exactly is used to program Ai-Da’s algorithms, but most generators like Stable Diffusion are being sued for using artists’ work without permission to train their own. According to Meller, “[the] vast majority of what Ai-Da does is… by what she’s seeing” and therefore not stealing the artist’s work.” So, a portion of the data she uses is camera input, but what about the “human inputs” that her website speaks of? This is likely referring simply to the fact that Ai-Da is not completely self-sufficient; she requires manual intervention to function as intended. Additionally vague is the type of artificial intelligence she uses, as well as how and when it gets updated. It can be assumed that the data for her AI, including the Large Language Model (LLM) behind her speech, is sourced from far and wide alike other AI programs. This suggests a contradiction in Meller’s statement—that Ai-Da is “not stealing” artists’ existing work—because he admits that camera input (“what she’s seeing”) only makes up the “vast majority of what Ai-Da does…” Therefore, if Ai-Da is as advanced and intelligent as is purported, how likely is it that she is given other data with permission from its sources? At best, the morality of this humanoid AI image generator—of physical images, that is—is unclear according to the information given by its creator.
Portrait by Ai-Da. Untitled, 2019, Ai-Da, 2019, https://www.ai-darobot.com/artwork.
Ultimately, Ai-Da is a more sophisticated, three-dimensional iteration of a combined LLM and image processor. But is she really an artist? Meller believes so; “Ai-Da is creative under the criteria set by Professor Margaret Boden, requiring works to be new, surprising and of cultural value.” This definition does seem to fit Ai-Da’s capability—she takes in information and generates something unique based on it. However, Meller makes a further claim that “Ai has agency. It can make decisions… [algorithms] can make creative decisions.” While AI does make decisions, it still lacks true sentience: the ability to feel. Even Ai-Da herself admits this: “‘I do not have feelings like humans do however I’m happy when people look at my work…’” What does it mean for an unfeeling thing to be “happy”, and can you be an artist without emotions? As for the art we produce, by human hand or AI generator, Ted Chiang believes the two are not interchangeable: “art requires making choices at every scale; the countless small-scale choices made during implementation are just as important to the final product as the few large-scale choices made during the conception.” He points to the limitations of AI decision-making at this time; image generators typically take in one initial prompt, lacking a human artist’s ability to fine tune along the way. Chiang further considers the likelihood that AI will soon allow this level of user input. He gives the example of film director Bennet Miller’s use of an image generator in this way: “he crafted detailed text prompts and then instructed DALL-E to revise and manipulate the generated images again and again. He generated more than a hundred thousand images to arrive at the twenty images…” Additionally, Miller was not able to replicate this fine tuning with a more recent version of DALL-E. Chiang concludes that traditional artists should not feel threatened by AI because these programs are not aiming to sell the capability of Miller’s laboriously successful result. Rather, AI companies “[want] to offer a product that generates images with little effort.”
Handwritten poem by Ai-Da. Untitled, May 31, 2022, Blooloop, May 31, 2022, https://blooloop.com/technology/in-depth/ai-da/.
Ai-Da and other up-and-coming AI “robot artists” are revolutionizing the meaning of art itself—once considered a strictly human practice. As machines and algorithms automate much of our labor, they are also targeting the creative domain. This may feel like a threat to artists and their work, and it certainly can hurt them through unauthorized data scraping. However, the possibility that increasingly human(oid) robots will replace human art is highly unlikely. Even if some believe these machines to be intelligent and creative, we still are separated by the ability to genuinely feel emotions. Art hinges on the feelings involved in its creation and feelings an audience gets from it. Chiang reassures, “We are all products of what has come before us, but it’s by living our lives in interaction with others that we bring meaning into the world.”
Bibliography
Ai-Da. “Ai-Da,” 2019. https://www.ai-darobot.com/.
“AI v the Mind: Meet the World’s First Artist Robot.” Uploaded by BBC, October 18, 2024. https://www.bbc.com/reel/video/p0jylq65/ai-v-the-mind-meet-the-world-s-first-artist-robot.
Chiang, Ted. “Why A.I. Isn’t Going to Make Art.” The New Yorker, August 31, 2024. https://www.newyorker.com/culture/the-weekend-essay/why-ai-isnt-going-to-make-art.
Flood, Alison. “Robot Artist to Perform AI Generated Poetry in Response to Dante.” The Guardian, November 26, 2021. https://www.theguardian.com/books/2021/nov/26/robot-artist-to-perform-ai-generated-poetry-in-response-to-dante.
Ho, Karen K. “Artwork Made by Humanoid Robot Ai-Da Using AI Algorithms Sells for $1 M. at Sotheby’s.” ARTnews.com, November 8, 2024. https://www.artnews.com/art-news/news/artwork-humanoid-robot-ai-da-artificial-intelligence-algorithms-sothebys-alan-turing-1234723391/.
Untitled. 2019. Ai-Da. https://www.ai-darobot.com/artwork.
Untitled. May 31, 2022. Blooloop. https://blooloop.com/technology/in-depth/ai-da/.
Hello Elaina, it is great to see that you were inspired to talk about Ai-Da from the relevant class presentation. Her existence alone poses many questions for the future of Ai art, and what it means for artists. A point you brought up that stuck out to me was the contradictions regarding her ability to feel, as it is stated she does not possess human emotions, yet describes being happy when people view her work. This may be to make the work more appealing to the general public by making her appear more human.
ReplyDeleteThis leads into a Ai-da being presented as more of a spectacle, where "the audience is encouraged to marvel and have fun playing a sort of 'guess the source' game, rather than critically pondering the social impact of AI technologies."[1] Interestingly, much else is left untold, as before a certain point the application of paint to canvas was not done by Ai-Da herself. Rather, "According to the labels accompanying Ai-Da’s paintings showcased at the robot’s first solo exhibition in Oxford in 2019, the coordinates from Ai-Da’s drawings were then “plotted onto the Cartesian plane and then run through AI algorithms." Then, the “digital versions of the artworks were transferred to canvas and overlaid with oil paint by artist Suzie Emery."[1] This brings into question if Ai-Da is truly created with intent to further the sphere of Ai art or if it is more so a stunt, designed to be marveled at.
[1] Lima, Eleonora. "AI art and public literacy: the miseducation of Ai-Da the robot." Springer. May 21, 2024. https://link.springer.com/article/10.1007/s43681-024-00488-5. Accessed November 22, 2024.