Mapping AI in independent cinema: aesthetic practices, production models, and sites of legitimation

This thesis examines how generative artificial intelligence is reshaping aesthetics and production models in contemporary independent cinema. The analysis centers on two underexplored territories: the short film as a privileged site for formal experimentation with AI, and film festivals as institutions mediating the legitimation of these emerging practices. Through comparative case study analysis, the research contrasts two paradigms of AI filmmaking. The corporate-sponsored model, exemplified by Eliza McNitt's Ancestra (produced with Google DeepMind) and the Tribeca-OpenAI partnership, prioritizes seamless integration, hyperrealistic coherence, and the suppression of algorithmic artifacts. The critical-independent model, represented by Andrea Gatopoulos's The Eggregores' Theory and the ecosystem surrounding Lago Film Fest and the Nouvelle Bug residency, foregrounds glitch aesthetics, algorithmic opacity, and sustained interrogation of the medium's conditions of production. The theoretical framework draws on international scholarship alongside Italian contributions that articulate a distinctive critical position: Gatopoulos's cyber-realist manifesto, HARIEL's theorization of hallucination as resistance to algorithmic syncretism, and Francesco D'Isa's philosophical writings on the ontology of AI-generated images. Ethnographic fieldwork conducted at Lago Film Fest 2025, including interviews with filmmakers and participant observation, supplements textual analysis with direct testimony from practitioners. The thesis argues that the future of AI cinema will be determined not by technological advancement but by human choices: which institutional models gain dominance, which aesthetic practices are legitimated, and whether generative systems ultimately serve creative exploration or corporate interests.