Perspectives from Practice: Algorithmic Decision-Making in Public Employment Services

Abstract

Algorithms are increasingly being implemented into core welfare areas as Public Employment Services. These data-driven technologies are implemented with the ambition to support caseworkers' decision-making capabilities, by profiling unemployed individual's risk of long-term unemployment. The research outlined in this paper investigates how we can study opaque technologies as algorithms from the perspective of the users (caseworkers) and those categorized (unemployed individuals) by these systems. This is done by combining established methods within Computer-Supported Cooperative Work, including ethnographic fieldwork and Participatory Design methods. I present preliminary results focused on caseworker's perception of the value of AI in job placement, and find documentation plays a central role in collaboration in casework. With this research, I am to contribute to a deeper understanding of how the organization of work is impacted by data-driven technologies like AI and explore ways to include the voice of unemployed individuals in the development of digital public services.