The Paradox of Anthroponormative Restriction: Artistic Artificial Intelligence and Literary Writing

Artificial intelligence, in the shape of stochastic machine learning models, has seen an increased use in artistic production in recent years. However, it makes an immense difference if such an ‘artistic Artificial Intelligence’ is conceived of as an autonomous agent or only as a tool in the context of a human-machine assemblage. In this paper, I introduce the distinction between a strong and a weak artistic AI, and suggest that each invites a specific aesthetic: The former is inherently anthropocentric, strives for the reduplication of existing artforms, and reproduces concepts of a postromantic tradition such as expression, genius, and creativity; it is anthroponormatively restrictive. The latter, on the other hand, allows for an experimental approach towards genuine artistic novelty unhampered by human models through, paradoxically, keeping a human in the loop. I illustrate this point by discussing Ahmed Elgammal’s ‘Creative Adversarial Network’ and the digital poetry of Allison Parrish and Zach Whalen.