Is 'AI Art' an Art?: Deleuzian Creativity on Generative Artificial Intelligence Images

Abstract

This paper examines the artistic status of Generative Artificial Intelligence (GenAI)-generated images, commonly known as “AI art.” The emergence of AI in art creation challenges conventional human-centric notions of artistry, raising a critical philosophical question: Can AI create art? With this, the paper distinguishes two types of GenAI-generated images based on their input-process-output generation. The Type-1 involves generating images based on precise prompts, relying solely on matching labels and descriptions through texts-to-images. These images are categorized as mere reproduction or imitations of and based on existing artworks. Contrastingly, Type-2, employing Extended Mind Thesis (EMT), suggests cognitive processes are not confined to the human brain but can extend to ‘technical artifacts’, like GenAI. This implies, AI serves as a cognitive artifact that augments human creativity and enables the creation of images which the human brain alone could not (yet) conceive. Justifying this distinction, Deleuze was used to argue that the replication and reformulation of these already-existing images imposes “images of thought” likewise recirculating these ‘clichés’ which Type-1 (re)enforces. However, the human-technology relationship shaped their endless becoming and co-evolution. Herewith, Deleuze's Assemblage as the process of spontaneous integration, disintegration, and recombination with heterogeneous multiplicities relates to EMT by arguing that cognition extends beyond the brain into the body, environment, and technological tools like AI, underscoring the importance of external elements in cognitive processes which augments artistic expression and human creativity. In this case, GenAI shall be thoroughly attuned to internal cognition so that they are part of the thought process in artistic creation. The paper concludes that Type-2 images, reflecting the extended mind’s generative capacity, qualify as art. However, Type-1 images, limited to mere imitation through its mechanical reproduction, lack the criteria of artistic creation. This distinction contributes to ongoing discussions on the intersection of AI technology, creativity, and the philosophy of art, as well as the implications of images perceived in the media in their capacity to influence its audience and society.