The culture of neural networks

The possibilities of generated cultural production have undergone fundamental changes in recent years, leading to a rethinking of existing approaches to the text and the artwork as such. To grasp this process, Zuzana Husárová and Karel Piorecký propose the term “neural network culture,” which captures a wide range of generative practices and reception mechanisms. The Culture of neural networks contextualizes the phenomenon of literary texts and other artifacts generated using the latest technological techniques. The generation of literary texts using neural networks is part of a broader cultural process, to which this publication formulates a position through the lens of literary science, media theory, and art theory.

The scholarly debate over this topic has been inconsistent—on the one hand, it underestimates the diachronic connections between generated texts and the tradition of experimental and conceptual literature; on the other hand, it does not sufficiently clarify the new-generation procedures and the contribution of human and technological actors in them. Therefore, Husárová and Piorecký propose the notion of synthetic textual art, which reflects the specific roles of the different actors involved in generative practice and its intermedial nature. In doing so, they approach the topic from both historical and theoretical perspectives, analyzing the current state of generative practice in all three basic literary types and in the intermedial space using selected foreign and Czech-Slovak projects. This state of affairs is often distorted in media discourse and even mythicized in terms of the capabilities of “artificial intelligence”; therefore, a critical analysis of this media discourse is essential. Finally, the authors summarize the implications of this stage in the development of generative practice on creativity theory and literary theory.