CRII: CHS: Humanizing Algorithms: Empirical and Design Investigations of Sensitive Algorithmic Encounters

University Of Colorado At Boulder, Boulder CO

Investigators

Abstract

This project will study algorithmic interactions and develop strategies for the human-centered design of systems that incorporate algorithms and their underlying data. Software developers of platforms of all kinds are creating features that make use of algorithmically curated content that leverage data about people's relationships, behavior, and identities. However, algorithms usually make decisions based on system metrics that are readily calculable, such as the number of likes, plays, and clicks. Even more sophisticated algorithms are limited by the social information explicitly given or inferred from provided data. As a result, algorithms can fail to capture the social context and human meaning that is important to the acceptability and success of the interactions these algorithms are meant to support. The research will investigate both algorithmic and human understandings of social data, especially when they diverge. By attending to divergence, the research can examine human expectations of algorithms, how misunderstandings might be reframed, and how subsequent action is informed by those divergences. Specifically, this project will identify (1) how people navigate sensitive algorithmic encounters; (2) how these encounters impact people; (3) what social concepts algorithms are failing to understand; and (4) what design strategies are needed to address sensitive content in algorithmic curation. To focus this work, the specific context of inquiry will be algorithmic encounters with content related to loss of life, given its prevalence and sensitivity at both communal and individual levels. The broader impacts of the work include: (1) developing guidelines around the curation of and interactions with social data related to loss of life, which can also be applied to other groups and experiences where algorithms should be sensitive; (2) demonstrating how designs that incorporate social data can adopt human-centered approaches to sensitize encounters with algorithmically curated content; (3) contributing to the development of design practices that encompass the design of interactions, systems, algorithms, and data; and (4) engaging students in multiple fields, including Information Science, Computer Science, Media Studies, and Communication through research and curricular activities focused on human-centered approaches to studying and designing social algorithms. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.