The Book of Sand

Aiiiii Art Center is pleased to present its inaugural exhibition The Book of Sand, featuring works from Certain Measures, Dabeiyuzhou, Entangled Others, Jake Elwes, Obvious, Sofia Crespo, and Yuqian Sun.

Aiiiii Art Center structures its future practice around annual research topics. Following the the theme of the 2021 International Conference on AI and Art held at Tongji University, Shanghai, May 20–21, 2021, The Book of Sand takes on the question “What is the author?”. Our manifesto, “Da(t)aism: A git-festo about AI and Art,” guides the direction of this exhibition, which involves a series of AI Art workshops.

Lines consist of an infinite number of points; planes an infinite number of lines; volumes an infinite number of planes, hypervolumes an infinite number of volumes… No, this, this more geometrico, is definitely not the best way to begin my tale.

The picture of an infinite line came to me, and I drew it: the line went on forever, right up to the edge of the universe. That is how the universe was born. [1]

The two sentences above begin our tale as well as an “imitation game” that might eventually displace personal intellectual enjoyment. We presuppose this text has been written by a human, but, in fact, the first sentence is extracted from The Book of Sand, by the Argentine writer Jorge Luis Borges in 1975, and the second was generated in 2021 by a GPT-2 language model based on the first sentence. In 1950, Alan Turing, who is widely considered to be the father of artificial intelligence, proposed the “imitation game,” also known as the Turing test, in his groundbreaking paper “Computing Machinery and Intelligence.” [2] Turing asserted that before the end of the twentieth century we would be able to program a computer with up to a billion bits of information, one capable mimicking human behavior so accurately that after multiple five-minute testing sessions more than thirty percent of human interrogators would be unsure if they were talking to a human or a machine. The machine would then be recognized to have passed the Turing Test and therefore be considered to possess an intelligence comparable to man’s. Today, as we approach the realization of Turing’s theory, the question he posed in 1950 is even more relevant: Can machines think?

The exhibition employs a literary imagination to juxtapose seemingly infinite, random generative art with Borges’s Book of Sand, a book that possess neither a beginning nor an end. The reader can turn the pages of the book but cannot predict the outcome, just as we cannot fully comprehend the operational logic of the “black box” embedded by and into a neural network. Against the joy of possessing the book, the fear that the book is not truly infinite grows, just as we fear that the infinite productivity of AI will render us captive, as well as the fear that AI’s “infinite creativity” is merely the outcome of smart permutations of existing human ideas. Borges describes an artificial intelligence-like non-technological medium—a book that could generate anything—that maps the beginning of posthumanism, one without technology, where the human subject is challenged by “new models of subjectivity” that emerge from our non-human counterparts. [3] 

But can machines think? Turing proposed that the answer depends on what a “machine” is and what “thinking” means. Are humans the only beings that are capable of thinking or creating? Do machines have subjectivity? Is subjectivity a requisite for thinking or creating? This attempts to subvert and transcend the traditional anthropocentric paradigm correlated to notions such as “genius,” “creativity,” and “autonomy.” Through the ages, human perceptions of “thinking” have been closely associated with intellectual activities that imply an intention, which is measured through action, regardless of if they are human-based or not. In this sense, can machines be considered to have intentions? Since intention is objectified by humans, our question on the possibility of machines thinking opens a scenario in which the main concern lies in the power mechanism between humans and our non-human counterparts: Is the artist in control of the machine? Or is the machine in control of the artist?

The Book of Sand is an exhibition made by the joint efforts of machines and humans. The output of the machines interferes with human curatorial and artistic practices as soon as we employ machine translation and language models. The exhibition hopes to build a poetic imaginary space through an enigmatic and fleeting constellation of AI artworks. When visitors enter Aiiiii Art Center, a former power plant, they become variables that influence the output through their interaction with the exhibition space. Centering around the non-anthropocentric idea of the “democracy of objects, in which rubbish is as important as human beings, the artists will show a variety of artworks generated using different AI techniques. [4] Their works exhibit contemporary changes in the means of artistic production and open up a debate about issues related to the ethical, social, and philosophical implications of artificial intelligence. We hope that through the displacement of notions such as “autonomy” and “creativity,” through the incarnation of creativity in the artificial, and through the disintegration of human subjectivity itself, our audience will come to ponder the true expanse of “thinking” and “intelligence” and grasp the very meaning of “creation” as conveyed by the insights of a non-human creator.

[…] between 0 and 1 lies an infinite number of values, images a finite number of pixels
[…] No, this concept of machine language is definitely not the best way to begin the tale.
“Look carefully, you will never see it again.” [5]

The Book of Sand is curated by Xi Li. Filippo Fabrocini and Kostas Terzidis of Tongji University College of Design and Innovation serve as academic hosts.