Tracing algorithmic imaginaries in everyday life (ESR1)

In the context of HCI, a central gap in the study of algorithms is the lack of theory and methods to approach their capacity to redirect change as a result of what algorithms are perceived to be, not what they are designed to be. Over the past year, my research has focused on developing a theoretical framework to support an ethnographic study of the mechanisms through which algorithms become part of material culture, and how shared perceptions of algorithms can lead to context adaptations. During an upcoming 12-weeks fieldwork, I will prototype diverse methods to support three conceptual shifts.

A) To measure the impact of algorithms in everyday life, we need to focus on the extraction and production of information, not data. This distinction matters because algorithms process data and produce information. For example, navigation algorithms correlate data to provide an output that modifies the information of the system they intervene in. This results in more people adopting the same information pattern, which inevitably leads to the loss of information in another part of the system–such as businesses and neighbourhoods losing value because of changes to mobility patterns–because the social and cultural practices that guide how people interact with algorithms tend to become flattened as the use-context becomes larger. Data describes linear change, but to understand the punctuated temporality of cumulative change, we need to focus on information.

B) To trace and map changes in the information of a system, we need to move from technocentric to ecological understandings of the system. From an ecosystem perspective, the agency that actors have in shaping the unfolding of a system by influencing social, cultural and physical adaptations is referred to as niche construction. A niche is a structural, temporal, and social context in which species coexist alongside hierarchies of selective pressures that guide behaviour and adaptation. While the more-than-human turn in HCI acknowledges the capacity of algorithms to be actants in the unfolding of social, cultural, and ecological practices, it does so with little attention to how algorithmic processes have become entangled with niche dynamics. Systems theory provides a boundary framework for tracing the passing and accumulation of information, but to explain how information leads to context adaptations, we need to understand the coevolutionary dynamics of information and physical systems from a niche perspective.

C) To understand how information and physical systems shape each other, we need to characterise the algorithm as a selective pressure mechanism. Selective pressure is a persistent condition in the environment where information is being exchanged that mediates the passing and adoption of information. A defining characteristic of selective pressures is that their position and influence in shaping the niche is contingent on how other actors in the system perceive, react and adapt to their presence. If the pressure that the algorithm can inflict in people’s lives is contingent on how algorithms are perceived, attention needs to shift from following the code to following the people, paying close attention to what people talk about when they talk about algorithms.

The research findings of the fieldwork will contribute novel theory to examine the social life of algorithms as it emerges in the intersection of perception and adaptation, and new interdisciplinary methods to trace and map perception and adaptation as a design material to understand change.