Rethinking The Documentary Pact in the Age of Generative AI: Transparency and the Ethics of AI in Documentary Production

Abstract

The present article examines how generative artificial intelligence reshapes the epistemological and ethical foundations of documentary filmmaking. Building on the concept of the documentary “pact” between filmmaker and audience, it argues that AI introduces a qualitative rupture by enabling the production of audiovisual material without an indexical link to reality. To address this shift, the paper proposes a five-level taxonomy of AI interventions, ranging from technical-assistive tools to fully generative, authorial systems. Through comparative case studies, it demonstrates that ethical evaluation cannot rely on binary distinctions between authentic and fabricated content, but must consider degrees of ontological transformation. The article advances the principle of calibrated disclosure, arguing that transparency should scale with the extent of AI intervention, and concludes that maintaining trust in documentary depends on aligning technological practices with clear and context-sensitive ethical communication.