Abstract
The article explores technology-human relations in a time of artificial intelligence (AI) and in the context of long-standing problems in social theory about agency, nonhumans, and autonomy. Most theorizations of AI are grounded in dualistic thinking and traditional views of technology, oversimplifying real-world settings. This article works to unfold modes of existence at play in AI/human relations. Materials from ethnographic fieldwork are used to highlight the significance of autonomy in AI/human relations. The analysis suggests that the idea of autonomy is a double-edged sword, showing that humans not only coordinate their perception of autonomy but also switch between registers by sometimes ascribing certain autonomous features to the AI system and in other situations denying the system such features. As a result, AI/human relations prove to be not so much determined by any ostensive delegation of tasks as by the way in which AI and humans engage with each other in practice. The article suggests a theory of relationality that redirects focus away from questions of agency towards questions of what it means to be in relations.