Algorithmic Interventions: The Power and Politics of Public Decision-Making Systems

Algorithmic interventions are a new and expanding avenue to influence social policy. Decision-making software is now being integrated within governing institutions ranging from the criminal justice system to social welfare programs. However, within any governing context, both the determinations about implementation and the software’s efficacy are dependent on many institutional and strategic constraints. In particular, I discuss three considerations when analyzing decision-making software used in government agencies or institutions: the intervention’s alignment, strength, and scope. The relative policy alignment of the algorithmic intervention focuses on whether an intervention seeks to strengthen a current policy or reform (or otherwise change) it. The strength of the algorithmic intervention refers to how binding the output is for the decision-making structure (i.e. requiring mandatory outcomes or providing a suggestion as a decision aid). Lastly, the scope of an algorithmic intervention refers to the number of points in the process that the algorithm impacts. In an apolitical world we might expect these considerations to be purely decided on bureaucratic optimization, but in the highly politicized and contentious arena of criminal justice or social welfare services, we can expect each of these decisions to be opportunities for strategic interactions between interested parties. This research explores these interactions and discusses how they might be considered when evaluating the implementation of such technology.