AI and Visual Trust
What does AI-generated material do to our trust in sound and images? Ethnographic films are the result of time-consuming productions, founded in anthropological method and theory. Participant observation in long-term fieldwork can extend over several years, with the anthropologist and camera present in various landscapes and situations. AI-generated images appear without the recording devices, but with clear prompts from humans and the limitations of machines. In the seminar «AI and Visual Trust» three researchers are presenting their work which all investigates, challenges and deconstruct audiovisual expressions made with AI.
Main content
The seminar is part of NAFA International Film Festival 2026.
9:00 Coffee
09:30–09:35: Welcome by Senior Curator Åshild S. F. Thorsen
09:35–9:45: Introduction by Carolina Némethy, General Secretary of NAFA.
09:45–10:15: Presentation 1 Haruna Inagaki
We artists and designers have been greatly affected by AI-hype, and it has severely shaken trust in various aspects of our lives. In this presentation, I will talk about my PhD project, "Metaphor of Power: Visual Storytelling to Question Absurdity in the AI Industry Under Capitalism," and introduce the creative resistance currently taking place in response to this situation.
Bio:
Haruna Inagaki is a Japanese filmmaker with a background in animation, illustration and art education. She is currently working on a PhD project in Artistic Research at the Institute of Design, KMD – Faculty of Fine Arts, Music and Design. Her project researches practical and artistic resistance against AI-hype by creating visual storytelling that uses metaphorical expressions to depict absurd situations. From an Asian feminist filmmaker's perspective, she aims to challenge and deconstruct the capitalist, patriarchal and colonial narratives surrounding AI.
10:15–10:45: Presentation 2 Gabriele de Seta
Can generative AI tools be used to create an ethnographic documentary? This talk describes an experiment that sought to answer this question by selecting a specific text-to-video model and attempting to generate enough footage to be edited into a documentary about China. From Chinese landscapes to Chinese cities, from local families to cultural performances, Latent China demonstrates how exactly AI models generate videos, where their properties come from, and what their limitations are when it comes to reproducing the documentary form.
Bio: Gabriele de Seta is, technically, a sociologist. He is a Researcher at the University of Bergen, where he leads the ALGOFOLK project (“Algorithmic folklore: The mutual shaping of vernacular creativity and automation”) funded by a Trond Mohn Foundation Starting Grant (2024-2028). Gabriele holds a PhD from the Hong Kong Polytechnic University and was a Postdoctoral Researcher at the Institute of Ethnology, Academia Sinica and at the University of Bergen, where he was part of the ERC-funded project “Machine Vision in Everyday Life”.
10:45–11:15: Presentation 3 Mihai Andrei Leaha
Sonic Memes, AI, and the Limits of Audio-Visual Trust.
A jingle for a candidate nobody has heard of appears on TikTok, shared thousands of times, sung by a voice too metallic to belong to anyone identifiable. Drawing on fieldwork with fact-checkers in Brazil and Romania and on the making of the ethnographic film A Matter of Facts co-directed with Roger Canals (Visual Trust Project), I propose the notion of the sonic meme, a melody that circulates across platforms and bodies and accumulates political charge as it goes. Generative AI now produces such melodies at almost zero cost and or traceable authorship, in styles tuned to region, genre, and class. What does disinformation sound like, and what can ethnographic film do with a melody that fact-checking cannot?
Bio: Mihai Andrei Leaha is Associate Professor in Visual and Multimodal Anthropology at UiT The Arctic University of Norway, where he teaches ethnographic filmmaking and multimodal practice at the MA programme in Visual and Multimodal Anthropology. He chairs the Commission on Visual Anthropology of the IUAES and co-edits ZARE: Horizons in Visual Anthropology. His research on visual trust, disinformation, and AI in Brazil, Spain and Romania was carried out within the ERC project Visual Trust and has appeared in Cultural Anthropology and Visual Studies, alongside earlier work on sound and embodied memory in the DIY electronic music scenes. He is currently completing the feature documentary A Matter of Facts, co-directed with Roger Canals.
11:30–12:00: Discussion
