When Oxford art gallery director Aidan Meller and his team created Ai-Da, a human-form robot that can draw and paint using “machine learning” algorithms, they sought to challenge commonly held ideas of what art is.1 Their website claims that the idea that “art is entirely a human affair” is no longer appropriate and that “Ai-Da creates art because art no longer has to be restrained by the requirement of human agency alone.”2 (They also claim that machines and algorithms are replacing human agency, an error I may address separately. Here I want to focus on their claims about art.)

Meller and his team’s achievement in designing and building a machine capable of creating intricate, original images and “learning” to do so in a range of styles is impressive. I commend them on their hard work and their contribution to the development of computer and robotics technology.

That said, their claims about Ai-Da’s creations evidence a fundamental misunderstanding of the nature and function of art.

Art is a means through which a creator communicates his ideas and values to viewers. As philosopher Ayn Rand observes, “Art is a selective re-creation of reality according to an artist’s metaphysical value judgments”—that is, his judgments about the fundamental nature of reality. “It conditions or stylizes man’s consciousness by conveying to him a certain way of looking at existence.”3 We can see this in the way that artists—be they painters, musicians, writers, or the like—use their art as a means to express their ideas and values, and in the way that viewers use art to expand their experience, understanding, and enjoyment of life.

Of course, at this point at least, robots are not conscious. They do not have values, make value judgments, or form or express views about the nature of the world. Robots do not select aspects of importance to convey to viewers, and—not being conscious or having perspectives on existence—they don’t consciously seek to convey perspectives on existence. Ai-Da does only what it is programmed to do. True, the robot’s programming makes it scan and compute new data, randomize it, and thereby create something Meller and his team did not preplan. But this doesn’t change the fact that Ai-Da is mechanistically acting as it was programmed to act, as does any computer-controlled device. Ai-Da is merely scanning data and mimicking human art as it mechanistically must, given its programming. It is not conceptually understanding art or selectively re-creating anything or being creative in any way.

One might argue that the images and objects Ai-Da makes are, in fact, Meller and his team’s artworks, made using Ai-Da as a tool. However, Ai-Da is not the kind of tool that aids an artist in selectively re-creating reality. Rather, it’s a tool for randomizing output and removing value-based selectivity—the essence of art—from the process and the outcome.

The composer John Cage similarly aspired to remove value-based selectivity from his work, resulting in such inanity as his “Imaginary Landscape No. 4 (March No. 2),” the “score” for which instructs the performer to tune twelve radios to specific frequencies throughout the piece. What the radios play depends on when and where the “piece” is “performed.” Cage designed this stunt to be “unimpeded by the service to any abstraction,” saying, “Value judgments are not in the nature of this work as regards either composition, performance, or listening.”

Ai-Da creates essentially the same kind of thing that Cage created: “artworks” devoid of artistic intent and, therefore, devoid of the essence of art.4

It may be that, someday, humans will learn how to create consciousness and develop a machine with this faculty. If so, perhaps the conscious machine will be able to make artworks of its own. But we should not celebrate that achievement before it happens.

Ai-Da by @AidanMeller is not the kind of tool that aids an artist in selectively re-creating reality. Rather, it’s a tool for randomizing output and removing value-based selectivity—the essence of art—from the process and the outcome.
Click To Tweet

Endnotes

1. Meller and his team use the term “artificial intelligence,” which is misleading. It is used nowadays to refer to machines that mimic the actions and abilities of humans but do not possess a sense of identity or the ability to think abstractly or form concepts. This is entirely different from the original conception of the term in science fiction to refer to a fully conscious artificial mind. “Machine learning” is also a little problematic as a machine cannot “learn” in the same sense as a conscious, thinking mind, but some can add new information to their programming and adjust it accordingly by themselves, and there isn’t currently a good alternative word to “learning” to describe this.

2. Ai-Da website, Aidan Meller Gallery, 2019, https://www.ai-darobot.com/artwork.

3. Ayn Rand, “Art and Cognition,” The Romantic Manifesto (New York: Signet, 1971), 35–36.

4. “Imaginary Landscape No. 4,” Art & Electronic Media Online Companion, https://artelectronicmedia.com/en/artwork/imaginary-landscape-no-4-2/ (accessed October 21, 2021).

Return to Top
You have loader more free article(s) this month   |   Already a subscriber? Log in

Thank you for reading
The Objective Standard

Enjoy unlimited access to The Objective Standard for less than $5 per month
See Options
  Already a subscriber? Log in

Pin It on Pinterest