算法排除：算法对稀疏和缺失数据的脆弱性（英）

Working Paper Algorithmic exclusion: The fragility of algorithms to sparse and missing data ______________________________________________________ Catherine Tucker This working paper is available online at: https://www.brookings.edu/center/center-on-regulation-and-markets/ The Center on Regulation and Markets at Brookings creates and promotes rigorous economic scholarship to inform regulatory policymaking, the regulatory process, and the efficient and equitable functioning of economic markets. The Center provides independent, non-partisan research on regulatory policy, applied broadly across microeconomic fields. February 2023 Disclosure Catherine Tucker has served as a consultant for numerous technology companies, a full list of which can be found here. The author did not receive financial support from any firm or person for this article or, other than the aforementioned, from any firm or person with a financial or political interest in this article. The author is not currently an officer, director, or board member of any organization with a financial or political interest in this article. Algorithmic Exclusion: The Fragility of Algorithms toSparse and Missing DataCatherine Tucker∗January 18, 2023AbstractThis paper introduces the idea of ‘algorithmic exclusion’ as a source of the persis-tence of inequality. Algorithmic exclusion refers to outcomes where people are excludedfrom algorithmic processing, meaning that the algorithm cannot make a predictionabout them. This occurs because the conditions that lead to societal inequality canalso lead to bad or missing data that renders algorithms unable to make successfulpredictions. This paper argues that algorithmic exclusion is widespread, and that itsconsequences are significant.∗Catherine Tucker is the Sloan Distinguished Professor of Management Science at MIT SloanSchoolofManagement.Shehasconsultedformanytechnologycompanies-pleaseseehttps://mitmgmtfaculty.mit.edu/cetucker/disclosure/21What is Algorithmic Exclusion?The age of algorithms is upon us. The digital era has allowed firms to collect and storeand parse data at far lower costs than ever before (Goldfarb and Tucker, 2019). However,in the past few years the biggest excitement has come from the idea of using algorithms ormachines to make better predictions using that data (Agrawal et al., 2016). Algorithms arebroadly used in digital advertising. In (Neumann et al., 2019), algorithms make predictionsabout whether someone who might see an ad is a man or a woman or interested in sports.Algorithms are used in education, where they assess the quality of instructors’ education(O’Neil, 2017). In our criminal justice systems, algorithms predict the likelihood that settingbail will be effective (Kleinberg et al., 2018). In HR, algorithms screen resumes to ensurethat recruiters focus on the best resumes (Cowgill and Tucker, 2017). Algorithms are used toidentify and allocate health spending (Obermeyer et al.