Classifying Contemporary AI Applications in Intermedia Theatre: Overview and Analysis of Some Cases

In the last years, several artists have begun to employ tools and software from the field of Artificial Intelligence for producing artworks. The most known collaborations on this topic concern the fields of visual art, drawing, plot writing, and music composition, but there are also several experimental uses of AI in theatrical performances. Here we propose a theoretical framework for contemporary theatrical pieces where AI becomes an integral part of staged actions. We study the relevance of AI as a non-deterministic element that fosters extemporaneous outputs, where the staging implies self-standing algorithms that affect dramaturgy and define peculiar artistic approaches. A cross-section of 13 works is considered to analyse the most recent applications and provide a comprehensive categorisation. Specifically, the framework entails two main phases of the artistic practice: 1) the preliminary setting of algorithms; 2) their function and representation on stage. The former regards the dataset definition and training process and highlights the author’s perspective in structuring the software for further staged performance; descriptions of the architectures of the algorithm are provided to delve into some implementations. The latter is related to the scenic interpretation of AI within the dramaturgical concept; examples of the mise-en-scène are considered to describe the role of the software in relation to human agents. The analysis proposes a rather broad and versatile preliminary model useful for both artistic and academic purposes that can be extended to future employments of AI.