Scholarlink

This chapter presents a theory of punitive bias in Immigration and Customs Enforcement (ICE). It introduces readers to the Punitive Optimization Problem (POP) in algorithm editing, which we define as the mechanism by which public officials adjust RCA scoring rules to minimize dissent by the officers charged with carrying out the immigration and detention policies. As officers dissent with “low-risk” more frequently than with “high-risk,” modifications to the RCA scoring rules result in harsher immigration rules over time.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Theory of Punitive Biases in Risk Assessment Algorithms

  • 章节

  • 发表日期：

    2025年

查看全文
  • Robert Koulish,
  • Ernesto Calvo

摘要

This chapter presents a theory of punitive bias in Immigration and Customs Enforcement (ICE). It introduces readers to the Punitive Optimization Problem (POP) in algorithm editing, which we define as the mechanism by which public officials adjust RCA scoring rules to minimize dissent by the officers charged with carrying out the immigration and detention policies. As officers dissent with “low-risk” more frequently than with “high-risk,” modifications to the RCA scoring rules result in harsher immigration rules over time.