Using Algorithms to Address Trade-Offs Inherent in Predicting Recidivism
Using Algorithms to Address Trade-Offs Inherent in Predicting Recidivism
复制标题
使用算法来解决预测累犯中固有的权衡问题
DOI：
--
复制
发表时间：
2020
影响因子：
1.4
通讯作者：
Christopher T. Lowenkamp
中科院分区：
文献类型：
--
作者：
Jennifer L Skeem;Christopher T. Lowenkamp
尽管风险评估越来越多地被用作帮助改革刑事司法系统的工具，但一些利益相关者坚决反对使用算法。主要的担忧是，通过安全地降低监禁率所获得的任何好处都将被算法本身所固有的种族正义成本所抵消。但是，公平权衡是预测累犯的任务所固有的，无论预测是由算法还是人类做出的。基于67，784名黑人和白色联邦监督者的匹配样本，通过定罪后风险评估进行评估，我们比较了“去偏见”算法的三种替代策略如何影响这些权衡，使用暴力犯罪逮捕作为标准。这些候选算法都能强烈预测暴力再犯罪（曲线下面积= 0.71-72），但与种族的相关性不同（r = 0.00-0.21），并改变了阳性预测值和假阳性率之间的平衡。为算法提供对种族的访问（而不是忽略种族或“掩盖”其影响）可以最大限度地提高校准并最小化不平衡的错误率。政策制定者与效率与公平的价值偏好的影响进行了讨论。
Although risk assessment has increasingly been used as a tool to help reform the criminal justice system, some stakeholders are adamantly opposed to using algorithms. The principal concern is that any benefits achieved by safely reducing rates of incarceration will be offset by costs to racial justice claimed to be inherent in the algorithms themselves. But fairness trade-offs are inherent to the task of predicting recidivism, whether the prediction is made by an algorithm or human. Based on a matched sample of 67,784 Black and White federal supervisees assessed with the Post Conviction Risk Assessment, we compared how three alternative strategies for "debiasing" algorithms affect these trade-offs, using arrest for a violent crime as the criterion . These candidate algorithms all strongly predict violent reoffending (areas under the curve = 0.71-72), but vary in their association with race (r = 0.00-0.21) and shift trade-offs between balance in positive predictive value and false-positive rates. Providing algorithms with access to race (rather than omitting race or "blinding" its effects) can maximize calibration and minimize imbalanced error rates. Implications for policymakers with value preferences for efficiency versus equity are discussed.