Will AI technology really kill creativity as some critics suggest? meta creativityat least in Eduardo Navas’ new book, The Rise of Metacreativity: The Aesthetics of AI After Remixing (2022). A must-read for anyone interested in the emerging relationship between AI and art, the development of these technologies should not be seen as a sudden turn-of-the-table phenomenon, but instead as a historical phenomenon rooted in us. and has the advantage of being placed within a theoretical framework. cultural, political and economic past. The purpose of the book, the authors say, is to demonstrate and examine “how advanced creativity has emerged and how it is linked to human history.” It is precisely this advanced state that he calls metacreativity.
As far as the visual arts are concerned, there are two most important issues that shape public discourse. First, can machine learning techniques be used to create art? Second, do their power come from the theft of other artists’ work, which is often present in the dataset? Both questions are irrelevant, and it’s good to see how Navas answers them without overstating their importance.
It’s a well-worn myth to think that a work of art is inextricably linked to the technical skill of its individual creator. The more I develop it, the more this link shows its habituation. Are cave paintings the result of the craftsman’s unique abilities, or are they part of a communal ritual? What about Eastern art, where imitation of past masters was often essential to maintain continuity of style? , collage, pop art, action painting, net art, conceptual art, collective art, and all the other practices that have liberated the work of art from the artist’s labor, technique, and even maker’s intention in various ways. . create it? The only standard that can be maintained seems to be that of Dino Formaggio, who wrote in 1973, “Art is everything that people call art.”
Navas is a reminder that art has long been separated from the labor of its creators. In the workshops of Renaissance and Baroque artists, we can already see the separation between art and manual labor. Here, apprentices worked under the guidance of masters and signed the finished works themselves. This method was later used in Andy Warhol’s factory. The most blatant example is undoubtedly Duchamp’s Readymade. This method proved to be very successful among later generations of artists, including Neo-Dadaists such as Robert Rauschenberg and Jasper Johns, as well as his 1970s pop his artists and conceptual his artists. has been proven.
Navas defines Chance as a “meta-algorithm” that can be used for specific actions. This is a technique used by Dadaists, Surrealists, Futurists in the first half of the 20th century, and later by multidisciplinary artists such as John Cage and Nam June Paik. The creation of art objects made of modular parts was creatively explored through collage art, where fragments of existing material were treated as modules that could be recombined to create new compositions. A close approximation to such a “prompt” in text-to-image software is a drawing installation by Sol LeWitt, performed by gallery or museum staff according to his written instructions (essentially an algorithm). . for example, Mural #715 (February 1993) read:
We’ll have to wait until machine learning is born, with the work of recent artists like the art collective Obvious, machines taking over nearly all the work, and preparing for the emergence of metacreativity. The author’s definition is “a cultural variable that emerges when the creative process goes beyond human production to include non-human systems.” According to Navas, this is a gradual process, a step further beyond postmodernism.
[W]Embracing post-modern and conceptual trends, the process of producing work can be delegated to shift to another meta order. There, artists do not create algorithms themselves, but instead computer programs. The creative process with parameters written by the artist.
Navas doesn’t discuss much about the data collection required to create machine learning software. We assume that these technologies have been around for some time, based on practices that are part of our cultural, political, and social issues (remixing, compression, modularity, etc.). Centuries of economic world. Getting them right is essential to addressing them and understanding the risks and opportunities that come with them.
But how do they work? These technologies are essentially algorithms trained on billions of bits of data. Even ignoring the human input required to make such software work, the impact of each object in the dataset on the output is utterly insignificant compared to a human-made artistic or creative work. Not important. The situation is different when machine learning is used for plagiarism (such as “draw Mickey Mouse”), in which case the responsibility lies with the user. These systems can only be developed using vast amounts of data and copyrighted material embedded in them. Once the algorithm is created, these data are no longer used. Those who created them claim that the process is protected by fair use laws, and in fact, individual contributions are negligible. However, there are companies that create their own software using such algorithms, which may seem unethical. That said, I believe these companies should produce open source software and public domain output. Datasets are common assets, and so are results (this is how Stable Diffusion really works). However, the discussion is still open.
At the heart of many of Navas’ insightful concepts is the notion of ‘remix’, which he borrows from the realm of music (a topic to which he devotes great chapters) and applies it to the realm of images and text. His Shock G quote of the musician illustrates this connection well.
It’s probably a little easier to get your hands on music than learning to play the guitar. truth. Just like it’s a little easier to take a picture than it is to actually paint. A photographer is to a painter what a modern producer, DJ, computer is to an instrumentalist, he is a musician.
What machine learning does is not a true remix, but the comparison provides a useful stepping stone as it never picks up the same elements as the material in the dataset it was built from. Navas does not mention software such as Midjourney, DALL.·E2, or Stable Diffusion, perhaps because their development was more recent than the completion of his book. But through Remix, we can easily understand the risks and possibilities of such software.
One of the main risks of machine learning is creating echo chambers. By relying entirely on algorithms to choose what content to experience, we live only within the biases that these systems have been trained to live with. True autonomy comes from “artificial general intelligence” (AGI), which is “an algorithm that can assign tasks to itself, rather than functioning within a range of actions for self-training written by a programmer”. belongs to These algorithms are still science fiction. But if they do appear in the future, their modes of creativity, which were truly inhuman at the time, may be inaccessible to us. The answer to is that even if they were created, only similarly formatted software can appreciate them.
Navas’ book is arguably an important step forward in the AI discussion, and the concept of metacreativity is a valuable tool. This highlights changes that appear to be quantitative rather than qualitative in nature. After all, even oil paintings like photography and computer graphics are subject to theoretical and technical knowledge, stylistic choices, limitations, and possibilities that have been practiced by many people over the years, centuries, and millennia. It incorporates a dense network of sexes.
Tools are not inert objects. The tool builds on the legacy of those who used, developed and modified it before us. In a way, every tool has its own will, which it must accept, because it internalizes and leaves ancient knowledge that is only revealed through its use.
Perhaps, to paraphrase Bruno Latour, we have always been metacreative.
¤
Francesco Disa is a philosopher and artist from Florence, Italy. He writes and draws for various magazines and is the editorial director of an Italian magazine. L’Indiscreto.
