Can sentencing be enhanced by the use of artificial intelligence
Can sentencing be enhanced by the use of artificial intelligence
复制标题
人工智能能否加强量刑
DOI：
--
复制
发表时间：
2017
期刊：
影响因子：
--
通讯作者：
M. Bagaric
中科院分区：
文献类型：
--
作者：
Nigel Stobbs;D. Hunter;M. Bagaric
人工智能正在重塑法律和法律实践的许多方面。然而，它现在才慢慢开始对更复杂的领域产生重大影响，这些领域涉及通过识别、排序和校准众多变量来做出决策。在本文中，我们研究了开发算法来做出量刑决定的可行性。尽管量刑决定受到 200 多种考虑因素的影响，但量刑法律和实践表面上适合自动决策，因为大多数相关事实是在相对较短的认罪听证会之前或之后确定的。此外，确定相关量刑目标以及加重和减轻处罚的考虑因素通常相对简单。我们建议应开发和试验量刑算法，作为现有量刑实践的辅助手段。如果试验成功，应考虑广泛使用计算机辅助量刑决定。
Artificial intelligence is reshaping many aspects of law and legal practice. However, it is only now slowly starting to have a major impact on the more complex domains, which involve reaching a decision through the identification, sorting and calibration of numerous variables. In this article, we examine the feasibility of developing algorithms to make sentencing determinations. Although sentencing decisions are influenced by more than 200 considerations, sentencing law and practice is on its face amenable to automated decision-making because most of the relevant facts are established prior to or following a relatively short plea hearing. Moreover, it is generally relatively straightforward to identify the relevant sentencing objectives and aggravating and mitigating considerations. We recommend that sentencing algorithms should be developed and trialed by way of being used as an adjunct to existing sentencing practices. If the trial is successful, consideration should be given to the wide-ranging use of computer-assisted sentencing decisions.