Beyond Algorithmic Scapegoating - Fostering Cultures of Algorithmic Responsibility Through Administrative Law and Design

Explore a keynote address from FAT* 2019 that delves into the concept of algorithmic responsibility through the lens of administrative law and design. Examine how speaker Deirdre Mulligan from the University of California, Berkeley, challenges the notion of algorithmic scapegoating and proposes fostering a culture of accountability. Investigate case studies involving gender bias, criminal justice, and urban planning to understand the implications of algorithmic decision-making. Analyze the role of policymaking, procurement processes, and legal frameworks in shaping responsible AI practices. Discover potential interventions, including the application of administrative law principles and the "arbitrary and capricious" standard. Consider the importance of expertise, contestability, and public participation in AI governance. Reflect on the challenges faced by legal practitioners and government agencies in addressing algorithmic bias. Gain insights into innovative approaches for improving public engagement in technology policy decisions.

Introduction
Background
Gender
Eric Loomis
Broadus Goodman
Assumptions
Operationalizing target variables
Choosing a model
Thresholds
Why is this happening
Robert Moses bridges
Procurement
Policymaking
Interventions
Case law
Administrative law
Arbitrary and capricious
What clause should agencies adopt
What are the options
Expertise
Contestability
Law Practice
No Voice
Inhouse
Government vs private sector
Policy choice
Eric Corbett
Using technology to improve public participation
How can the public be prepared