In the contemporary media landscape dominated by discourse surrounding AI-generated content, this project inverts the paradigm by examining AI not as a creator, but as a spectator of human-authored narratives. This research explores the profound implications of developing systems capable of consuming and responding to human creative expressions, particularly within the realm of cinema.
The investigation centers on a novel multi-modal AI system designed to simulate emotional responses to filmic stimuli through synthesized internal dialogues and automated facial expressions. This system serves as both observer and subject, creating a complex feedback loop of spectatorship that challenges traditional notions of audience, perception, and emotional authenticity. By generating synthetic reactions based on learned patterns of human emotional response, the system raises fundamental questions about the nature of media production and consumption.
The research culminates in a series of nested observational experiments: AI reacting to cinema, humans reacting to AI's reactions to cinema, AI reacting to humans reacting to AI's reactions to cinema, and subsequent recursive iterations of this pattern. These layered interactions create "Mise en abyme", a network of spectatorship that mirrors the increasingly complex media ecology of our time. Through this framework, the project examines how the traditional boundaries between observer and observed dissolve within our cybernetic society, where the act of watching and being watched exists in perpetual recursive dialogue. By positioning AI as a spectator rather than a creator, the project offers new perspectives on questions of authenticity, simulation, and the nature of emotional response in an age where the lines between natural and synthetic experience become increasingly blurred.