Human-Algorithm Relationships: Moving Beyond the Interaction as a Site of Empirical Research

As the algorithmic retrofitting of everyday interactions steadily advances, so does the need to understand how humans and algorithms influence each other over time. For this purpose, the interaction has become the primary site of empirical research. Yet, the relational, spatial and temporal boundaries that the interaction brings to bear are poorly defined, largely outdated and often restrictive. This results in novel findings aggregating to already worn-out narratives where the interaction is the only way algorithms participate and transform everyday life. In contrast, the concept of relationship, which entails relational, emotional, temporal, and spatial dynamics more attuned to how humans and algorithms increasingly influence each other, has received little attention. Borrowing Robert Hinde's framework of "patterned interactions=relationship,"this provocation challenges the limits of the interaction and argues that the relational boundaries of a relationship open up new opportunities for empirically researching how humans and algorithms influence each other over time.