Emerging evidence from research on algorithmic decision-making in the public sector suggests that AI systems are leading to outcome-oriented harms that disproportionately impact low-income and minority communities. Here, the interplay between the legal and systemic mechanics and AI systems can adversely impact the fairness of the decision-making process itself. I unpack process-oriented harms in child welfare that adversely affect the nature of professional practice, and administration at the agency, and lead to unreliable decisions at the street level. Caseworkers are compelled to undertake additional labor in the form of repair work to restore disrupted administrative workflow and decision-making processes, all while facing organizational pressures and time and resource constraints.
Resources for Attendees
Devansh Saxena is a Presidential Postdoctoral Fellow at Carnegie Mellon University and will be joining the Information School at the University of Wisconsin-Madison as an Assistant Professor in Fall 2024. He studies sociotechnical practices of decision-making in high-stakes domains and the social impacts of introducing AI in such domains. His work examines how human-AI interaction plays out in practice where decisions are mediated by organizational constraints, nuances of professional practice, and algorithmic decision-making. He is interested in developing new methods and tools that support AI innovation at the earliest stages of ideation, problem formulation, and project selection.