Working with Algorithmic Colleagues: Expectations and Experiences of Automated Decision-Making
- Contact person:
- Martin Berg
- Swedish Research Council
- Responsible at MaU:
- Martin Berg
- Time frame:
- 01 January 2021 - 31 December 2024
- Research environment :
Technologies of automation are increasingly transforming organisations across the world. Automation is no longer exclusively about making things; it frequently involves making decisions. Automated Decision-Making (ADM) is implemented in organisations to foster efficiency, productivity, and precision. Lately, instances of automation in public administration have attracted considerable attention in Sweden.
Roughly twenty municipalities have implemented ADM, and critics have raised their voices against ADM as a threat to professional practice. Although studies have pointed out how these changes might challenge public administration, no research exists that engages with the social dynamics of ADM implementation, from the design of ADM to policymaking and management of their integration, to how it feels to work with ADM as an ’algorithmic colleague.’
With an initial focus on ADM for income support in three regions, this ethnographic project responds to this gap by answering the following research questions:
- How do developers at IT companies contracted by public administrative bodies expect ADM to support or even replace work tasks in public administration?
- How do policymakers and managers expect to integrate current and future ADM in the organisational structure and institutional fabric of public administration?
- How do professionals in public administration experience and interact with current ADM, and how do they imagine work with future ADM? Aiming at understanding ADM as utterly social, in its design, implementation and consequences, this project produces essential knowledge about the development toward a 'smart' public administration.
Putting power relations and potential friction between stakeholders on different levels under scrutiny, this project opens up additional ways to think of policies, practice and design, which account for how humans are inevitably co- implicated with the systems, processes, effects and affects of ADM.