Struggling with generative AI: Digital self-determination along infrastructures of automation

Mollen, Anne

Abstract

The widespread adoption of generative artificial intelligence (AI) has led to struggles on a global scale by individual and collective actors trying to secure their autonomy with reference to generative AI. There are several examples of how generative AI impacts the ability of individuals and collectives to self-govern and exercise their free will: for instance, training data copyright violations, cultural misrepresentations, precarious working conditions of data workers, and the environmental and social justice implications of generative AI development. Based on discussions from various fields, including machine learning, AI ethics, and critical data studies, this article presents how current struggles in generative AI relate to matters of autonomy, sovereignty, and self-determination. It specifically reflects on autonomy in relation to generative AI training data, accountability, and market concentration as well as social and environmental justice. Given that these struggles over autonomy significantly relate to the materiality of generative AI, the article proposes digital self-determination and an infrastructural perspective as an analytical concept for a multi-actor, process-oriented, and situated contextual analysis of how autonomy implications manifest in generative AI infrastructures.