ABSTRACT
Why do biased algorithmic predictions arise, and what interventions can prevent them? We examine this topic with a field experiment about using machine learning to predict human capital. We randomly assign approximately 400 AI engineers to develop software under different experimental conditions to predict standardized test scores of OECD residents. We then assess the resulting predictive algorithms using the realized test performances, and through randomized audit-like manipulations of algorithmic inputs. We also used the diversity of our subject population to measure whether demographically non-traditional engineers were more likely to notice and reduce algorithmic bias, and whether algorithmic prediction errors are correlated within programmer demographic groups. This document describes our experimental design and motivation; the full results of our experiment are available at https://ssrn.com/abstract=3615404.
- Amanda Agan, Bo Cowgill, and Laura Gee. 2019. The Effects of Salary History Bans: Evidence from a Field Experiment. Working paper (2019).Google Scholar
- Marianne Bertrand and Sendhil Mullainathan. 2004. Are Emily and Greg more employable than Lakisha and Jamal? A field experiment on labor market discrimination. American economic review, Vol. 94, 4 (2004), 991--1013.Google Scholar
- Nitesh V Chawla, Kevin W Bowyer, Lawrence O Hall, and W Philip Kegelmeyer. 2002. SMOTE: synthetic minority over-sampling technique. Journal of artificial intelligence research, Vol. 16 (2002), 321--357.Google ScholarCross Ref
- Corinna Cortes, Mehryar Mohri, Michael Riley, and Afshin Rostamizadeh. 2008. Sample selection bias correction theory. In International conference on algorithmic learning theory. Springer, 38--53.Google ScholarDigital Library
- Bo Cowgill. 2019. Automating Judgement and Decision making: Theory and Evidence from Résumé Screening. Working paper (2019).Google Scholar
- Bo Cowgill, Fabrizio Dell'Acqua, and Sandra Matz. 2020. The Managerial Effects of Algorithmic Fairness Activism. In AEA Papers and Proceedings, Vol. 110. 85--90.Google ScholarCross Ref
- Bo Cowgill and Patryk Perkowski. 2019. Agency and Homophily: Evidence from a Two-Sided Audit. Working Paper (2019).Google Scholar
- Bo Cowgill and Catherine E Tucker. 2019. Economics, fairness and algorithmic bias. preparation for: Journal of Economic Perspectives (2019).Google Scholar
- Jon Kleinberg, Himabindu Lakkaraju, Jure Leskovec, Jens Ludwig, and Sendhil Mullainathan. 2017. Human decisions and machine predictions. Technical Report. National Bureau of Economic Research.Google Scholar
- Manish Raghavan, Solon Barocas, Jon Kleinberg, and Karen Levy. 2020. Mitigating bias in algorithmic hiring: Evaluating claims and practices. In Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency. 469--481.Google ScholarDigital Library
- Andreas Schleicher. 2008. PIAAC: A new strategy for assessing adult competencies. International Review of Education, Vol. 54, 5--6 (2008), 627--650.Google ScholarCross Ref
- Bianca Zadrozny, John Langford, and Naoki Abe. 2003. Cost-Sensitive Learning by Cost-Proportionate Example Weighting.. In ICDM, Vol. 3. 435.Google Scholar
Index Terms
- Biased Programmers? Or Biased Data? A Field Experiment in Operationalizing AI Ethics
Recommendations
A preliminary study of crowding with biased crossover
GECCO '13 Companion: Proceedings of the 15th annual conference companion on Genetic and evolutionary computationThis paper proposes a novel crowding method, which is called "Crowding with Biased Crossover (CBX)". The Biased crossover operator begins with two parents. Then two offspring individuals are created, each offspring taking more characteristics from one ...
Biased parameter adaptation in differential evolution
AbstractThe paper describes the problem of bias in parameter adaptation in Differential Evolution and other Evolutionary Algorithms. Based on the newly proposed Expected Fitness Improvement metric, the shift towards exploitation is ...
Design, Conduct and Analysis of a Biased Voting Experiment on Human Behavior
International Workshop on Behavior and Social Informatics on Behavior and Social Computing - Volume 8178To explore whether and how different social activities and phenomena, network structures and incentive machanisms could influent human behavior in social networks, Prof. Kearns and his colleagues conducted a series of human participated behavior ...
Comments