Mediating Conflicting Notions of Fairness

Abstract

We expand on previous work at Aequitas, and build upon existing works on designing worksheet and checklist for stakeholders of algorithmic systems, by developing a checklist for policymakers to map their notion of fairness onto metrics that are commonly used. We use the checklist as a conversational tool to workshop and examine where the understandings of fairness diverge across different human services domains and stakeholders (community member vs policymakers). This tool was presented at APPAM workshop. You can checkout the tutorial here. Advised by Prof. Rayid Ghani.

Lingwei Cheng

PhD in Public Policy and Management

I’m interested in developing and evaluating the effective and responsible applications of AI and data-driven tools in social policy areas including child welfare and homelessness.