ABSTRACT
To claim that AI (artificial intelligence) cannot exhibit creative behavior carries a weight that forces us to be genuinely open to the possibility that those making such a claim must have an implicit assumption or even a reliable and explainable account of what creative behavior is. Yet such accounts offer bizarre but familiar formulations of creativity as spontaneity that certain myths of modernism would have us believe are not able to be mapped onto any intelligible system, because, as with Kant, were such behaviors intelligible, they might also have to be capable of being taught and learned, and learning was for Kant, curiously, merely imitation. Yet the computational machine is now performing, or at least apeing, these procedures of the imagination hitherto ascribed to the secretive technics of nature. This chapter will argue for why such antediluvian notions hailing from an eighteenth-century worldview encapsulated in Kant’s transcendental project are all still exceptionally enmeshed in contemporary art’s critical paradigm of post-conceptual practices and research-based artistic practices, and more significantly, in general culture in ways that contribute to anxieties about the automation of creativity as a threat to livelihood, to intellectual property rights, and to the imagination itself. At the opposite end of such anxieties are the millions engaged in constructing a pandemonium of mimicked and generic artistic subjectivities tethered to their newly realized expressive capacities afforded by the democratization of the means of imaginative production encoded into an algorithm.
It will be shown that herein lies an exceptionally sticky historical situation from which we must work to extract ourselves. In order to adequately approach the possibilities of doing so, this chapter will (1) provide a vivisection of the notion of creativity itself, in its historical and contemporary determinations as a quite recent conceptual apparatus defining deviations of human behavior from drudgery, which will then (2) allow us to depict its contemporary ideological manifestation as historically embedded in Kant’s transcendental project, with some familiar art-historical signposts that will then (3) be shown to lead to an understanding of the new metaphysics of nature opened up by Kant, from which are derived a panoply of concepts (such as that of ‘emergence’ or ‘opacity’) that functionally align with the advantages and opportunities of deception, mimicry, and deviant behavior derived from biological history also informing contemporary practices of various kinds, including their foundational import to the computer. This will (4) lead us to an attempted reconciliation with contemporary art’s critical paradigm by elaborating how practices might benefit from considering the cognitive implications of an artificial implementation of pattern recognition abilities in machines retrieved as both perceptual and conceptual method that, alongside heuristics and other more recently proposed methodologies, can help us to defease the complexities just described. Following from this, an attempt will be made to extract some perspectives from domains such as ecological psychology, the notion of Dennettian ‘design space’, or the constructibility of models and theories that, moving forward, might enable us to adequately adjust our relations to these new AI models for practices be they theoretical, artistic, or otherwise.
