FRArt Latent Walk, beyond the black box.

1. Introduction

  • type: Article
  • ref: DOC.2021.147

Since the industrial revolution, the relationship to machines has entered societies and has taken an important place in public debates. Then automation came as a challenge to humans position in the world. These questions, which were theorized since the birth of cybernetics at the end of the 1940s, are not new, but they actually take on a whole new dimension today with the development of "artificial intelligences" which have succeeded in occupying a place in all dimensions of our society.
What we group together under the banner of AI (or Machine Learning) are these processes through which machines manage to learn on their own, on the basis of data corpus and neural networks. One could say that machines are empowering themselves and that, in this sense, they are no longer exclusively tools. If human intervention decreases in this process, I nonetheless tend to note, on the contrary, that the mystery and extent of their calculations are actually growing. In the media, in cultural production, in literature (and even scientific literature) dealing with AIs, the word “opaque” is often used to qualify the operations performed by algorithms. However, as an artist, researcher or teacher, it is our duty to go beyond both the fascination and the fear that they can arouse in order to think about the cultural changes they generate.

Taking at face value the belief that states that, in AI, the machine is autonomous (and therefore has, in a way, free will), my research will aim to show the unconscious of this machine, to make visible what is hidden inside his brain, his black box. My research will materialize through the creation of a functional artefact, a sort of 21st century "camera obscura", both as an allegory and as a tool, as a space of technological and aesthetical convergence.
On the one hand, the materialized object will incorporate an interactive installation and, on the other hand, a film in production. The interactive installation will allow each user to connect and collaborate with Artificial Intelligence in order to create an image. As for the film, it will stage the object created in an experimental fiction in order to question, through other means, the limits of the gaze and the relationship to the image in the context of the visual upheaval generated by artificial intelligences. These machines have now the capacity to produce a completely false "realistic" image (OR invented). By working to push these algorithms to their limits, by taking them by default, by searching for their failures, I want to question their announced sovereignty, and show that they can once again become simple tools, and therefore creative levers which artists can seize (just like, contradicting Walter Benjamin, they seized the camera).

Through my background in visual arts, it is particularly the ability of AIs to produce images that I wish to study and to question. Thus, the created object must be capable, previously trained by a team of researchers, of producing hallucinated images autonomously.

First, my research will consist in recontextualizing the artist / machine relationship in order to place this new technology in a historical continuity. Then, associated with artisans and object designers, it will be about materializing the artifact that will visually and plastically embody the neural networks of Artificial Intelligence. This black box will be at the center of an interactive installation.

For the IT part, the research will be supported by UMONS researchers via the CLICK. For the conceptualization of the object, a partnership with the TRAKK of Namur is envisaged as well as a collaboration with the ESA Saint Luc Tournai (and more particularly the "object styling" option). In addition, the results of this research will be shared with the students and will be the subject of a publication.