Digital images generated by Artificial Intelligence as ethnographic experimentation

Abstract

The article discusses the role of images produced by artificial intelligence (AI) in visual anthropology, highlighting their ability to represent identity and experience. It also addresses the technical and ethical challenges of AI data classification and its interaction with socio-technical networks, questioning technological neutrality. Technical implications include data categorization that can perpetuate biases and power relations. The simplification and distortion of representations by AI is highlighted, requiring a critical analysis of the stories embedded in the categorizations. It is proposed that anthropologists examine the relationship between image, label, and referent, recognizing differences and similarities in their roles and those of AI labelers in knowledge production. In addition, it discusses how AI images can influence anthropological interpretation and analysis by blending reality and emotion. It is argued that a critical engagement with the ethical and technical implications of the generation and use of these images is necessary.