2023/11/11
 11/17 政策科學資料分析全球線上論壇 ~2023系列VI講座：Legal Professionals' Attitudes Towards Algorithmic and Human Decision-Making


 
      REGISTER HERE     

Abstract: Decisions in criminal cases are influenced by the advice received from a multitude of legal experts and increasingly from algorithms or AI. While scholars and practitioners have raised concerns about the bias of algorithms in the criminal justice field affecting hundreds of thousands of defendants a year, few studies focus on how the legal community itself views advice from algorithms as compared to humans. Such a lacunae in the literature is problematic as it assumes that trained legal professionals, poised to influence algorithm development and use, have the same trust or distrust in algorithms as the general public.  To address this, we conducted three survey experiments from a unique national sample of legal professionals comparing trust in advice received from algorithms to that from humans.  Paradoxically, our results suggest that trained legal professionals or experts do not experience the “algorithms aversion” suggested in much of the social science literature on algorithm trust, despite general concerns of the justness and fairness of using algorithms in criminal cases. The methods used in this paper include multi-level models, parallel regression for estimating moderator effects, and traditional regression. The use of covariate balanced propensity scores and causal forests are demonstrated.

Bio: Ryan Kennedy (Ph.D., The Ohio State University, 2008) is a professor in the department of political science at the University of Houston (UH), principal investigator for the NSF-funded Community Responsive Algorithms for Social Accountability (CRASA) project, director of the Machine-Assisted Human Decision-making (MAHD) Lab (UH), founding associate director of analytics for the Initiative for Sustainable Energy Policy, editor of Research & Politics, and a research associate at the Hobby Center for Public Policy.