Spectral and Procedural: A Perspective on Artificial Creativity Through Computational Art

Rosemary Lee and I will present our paper Spectral and Procedural: A Perspective on Artificial Creativity Through Computational Art at the Conference on Artificial Creativity organised by the Malmö University.

This paper looks at the computational and at computational art to understand creativity: how actions and outputs of computational systems can be interpreted as creative by their human counterparts (programmers, interactants, readers, etc.). The paper starts by framing computational artworks as those that resort to computation in their realisation, whether or not digital computers are directly used or if computation is instantiated by other means. Grounded in conceptual art and in generative approaches to art, it then frames computational art as both infrathin and hyperobjectual, exploring how its traits are conducive to the development of radically new aesthetic experiences. If before the computational we had mechanical reproductions and replicas, we now have individualised simulations that manifest unique and idiosyncratic behaviours and developments. We find that if the computational draws on the other arts and furnishes them with new codes and techniques, it does not so much extend those other arts as it breaks from them in a mutation, an emergence of representation that takes steps beyond representation. Although the computational manifests sensorially, it is always non-perceptual. It requires a substrate, but computation is not to be found at that substrate, rather it exists between it and its future. The computational is not at an object’s perceptual surface, it manifests through the surface, but is found in the processes within and beyond its object. From the viewpoint of computer science, the computational is always deterministic, a trait that may seem to counter its potential creativity. The computational is also often irreducible and therefore, although deterministic, it is in practice impossible to anticipate its outcomes. Furthermore, the computational is increasingly complex, with systems easily becoming impenetrable blackboxes, not only for their users but even for their programmers. This leads to an analysis of computational arts as spectral phenomena, standing between an algorithmic past (a known and predictable field of formulaic and predictable repetition) and a futural being (a phase-space created by computation while it constructs itself through its execution). The futurality of the computational and of computational art is deeply connected to their irreducibility and to the possibility of being understood as autonomously creative, i.e. as not merely encoding human creativity (in the past) but rather manifesting creative agency (in the future). The artefacts in computational art retreat, they subscend as computation executes, because it is the execution that builds the aesthetic experience, leaving behind only artefactual remains. This, the paper argues, is the essence of artificial creativity, a process, not its outcomes.