Governing (with) algorithms: trust, uncertainty and the promise of transparency

• In Event: Artificial Intelligence and Cognition as Social Praxis

Abstract

In the risk regime, there is an international trend to use automated systems to support administrative and technical decision-making processes with the promise of achieving more efficiency and objectivity. However, a growing literature has described a number of algorithmic harms deriving from information and secretive decision-making scenarios or black-boxes. STS studies have focused on understanding algorithms’ harms, opacities, and regulations. Yet, the way how political actors have discussed regulations for these technologies seems unexplored. How has transparency been practiced by governmental actors? How has the control or access to information been promoted? Our argument is that transparency is in a tension between uncertainty and public trust, where the institutional openness performs the citizen agency, making the transparency a legitimate political practice. We present a two-part critical analysis of the NY City Council Bill No. 1696 that proposes to regulate and demand information from city agencies that use algorithms to deliver public services. Based on Burrell's (2016) framework of opacity, we compare the content of the bill with European Union’s prior regulation. Second, using the Committee on Technology’s last discussion meeting recording, we present its favorable arguments and interested parties reactions and concerns. Finally, three main reflections are presented about (1) trust and transparency in the risk society, (2) the need and complexity of studying the regulation of algorithms in practice and with real political stakeholders, and (3) the effects of the NY case on cities adopting smart planning trends and how IT creates uncertainty about government autonomy and citizen agency.

Authors

• Pedro Pablo Seguel, University of Texas at Austin

• Rodrigo Gonzalez, Independent