Directing the Algorithmic Gaze: A Framework for Machine-Readable Films

Bridging the Dual Audience: Aligning Human Intent with Algorithmic Logic The contemporary film industry faces a dual systemic crisis: an architectural inequality rendering films invisible to algorithms due to legacy workflows that result in creative outputs as black boxes or unusable data, severely limiting their discoverability and structural integrity, and a production barrier gated by platform monetization models and automated distribution patterns. This dynamic leaves independent creators vulnerable to the "Algorithmic Gaze" - an external system that misinterprets, curates, and normalizes their work. Within this landscape, cinema serves as a highly complex laboratory—an art form that inherently encompasses text, image, sound, and temporal editing, making it the ultimate testing ground for narrative systems (with direct scalability to the broader creative industries and digital media sectors). For independent and marginalized voices, this structural gap creates a triple vulnerability—cultural, economic, and algorithmic. This methodology addresses the fundamental challenge of content that is born structurally invisible to digital infrastructures, aiming at translating abstract narratives into executable operational layers. In the context of this research, a machine-readable film is not a different kind of film, but a film that behaves differently. From Form-First to Meaning-First for Enhanced Visibility This presentation—developed as an integral part of my doctoral thesis in Cinema and Media at UNATC—proposes a paradigm shift to Meaning-First Cinema. In the context of the Semantic Web, this means structuring sense not only for human audiences, but also for algorithmic processes. Responding to the industry's post-factum patchwork solutions for digital visibility (such as reactive data tagging, content enrichment, or traditional SEO and marketing), this framework bridges cinematic design with Information Architecture (IA), fundamental metadata, AI, and Semantic Web principles. By approaching the film as a self-communicating narrative system, it enables the creator to utilize the work’s native ontology and artistic vision as an "Aesthetic Operating System" to ensure core intelligibility for human, AI, and algorithmic audiences alike. As a filmmaker, I approach this issue through the lens of artistic practice and the realities of self-distribution. While introducing a framework for the digital and algorithmic literacy of the cinematic work itself—which inherently fosters the digital literacy of the filmmakers, film producers and distributors—this approach is specifically tailored to fit their workflows and to address the present gaps in the traditional supply-chain. It may also prove useful for technical specialists or developers interested in the specifics of cinema, helping them build applications or workflows that better support the creative process. From Pixel Manipulation to Semantic Design: Structuring the Creative Vision For demonstration, the methodology utilizes an applied workflow where an AI-augmented short film is constructed via an Informational Montage process, proving that shifting focus from pixel manipulation to semantic precision bridges the gap between human intent and machine response. Crucially, it reveals that the creator’s structured semantics act as superior "sources of truth" compared to traditional distribution pathways or empirical prompt engineering. By encoding cinematic vision into an architecture that is easily translatable into structured data, filmmakers achieve a multimodal objective: story and communication resonance, production efficiency and organic discoverability. This methodology highlights a new dimension for creative leadership—directing not just a crew or a project, but the underlying digital infrastructure itself, transforming the film from an isolated destination into a transparent node wired directly into the network. The final cinematic outcome demonstrating this architecture can be accessed here: Watch "San Pedro’s Dream" (2026, 3 minutes, AI-augmented short film). Dissemination and Dual-Track Validation This transdisciplinary research has been successfully peer-reviewed and disseminated across dual tracks: National Track (In Romanian): Presented at CINETic during the institutional research seminar "Stimularea și diseminarea rezultatelor cercetării în cadrul UNATC 2025" (Stimulating and disseminating research results within UNATC 2025). International Track (In English): Presented at the international hybrid conference "Women in Film: Leadership, Creativity, and Interdisciplinary Perspectives" (organized by Creatrix Fama and UNATC). Note: This repository contains the original presentation assets in both English and Romanian.