Algorithmic Memory and Epistemic Transformation in Documentary Cinema: An Analysis of Artificial Intelligence, Representation and Women’s Labor

Abstract

This study examines how generative artificial intelligence (AI) transforms reality, evidence, and testimony in documentary cinema. It treats the 68-minute documentary Tütün Kokan Kadınlar (Women Who Smell of Tobacco) (Kuzucan, 2023), based on interviews with 22 women who worked in Gerze’s tobacco fields or at the Gerze Tobacco Processing House, as a qualitative single case. AI-generated sequences totaling 4 minutes and 13 seconds are examined through visual-discursive analysis. The findings show that synthetic images provide visual continuity for oral testimonies by addressing archival gaps, while also organizing women’s labor through recurring codes of body, dress, and affect. Production scenes emphasize discipline, fatigue, and collectivity; care and family scenes foreground affection, happiness, and tradition. Increased visibility, therefore, coexists with a reduction of individual experience to recognizable cultural types. Conceptualizing this process as “algorithmic memory,” the study argues that AI reconstructs rather than records the past. Its responsible use in documentary requires transparent disclosure, preservation of testimony’s epistemic priority, and critical scrutiny of gendered visual conventions.