Sentencing by Machine: The Ethical and Constitutional Limits of AlgorithmicRisk Assessments

Abstract

The integration of artificial intelligence into judicial decision-making, particularly through algorithmic risk-assessment tools, has redefined the boundaries of sentencing discretion and constitutional guarantees. While algorithmic systems promise greater consistency and objectivity, they often reproduce and solidify existing inequities through opaque processes and immutable classifications. This article investigates the constitutional and ethical limits of algorithmic sentencing, analyzing how algorithmic immutability—the tendency of predictive systems to freeze individuals into fixed categories—creates new forms of discrimination that circumvent traditional equal-protection doctrines. Drawing upon comparative jurisprudence in the United States and the European Union, the article argues that algorithmic risk assessment systems challenge fundamental legal principles such as due process, transparency, and fairness. By exposing the structural dangers of delegating sentencing discretion to data-driven instruments, the paper advances a framework for doctrinal reform that emphasizes transparency, procedural fairness, and accountability. It concludes that without strict regulation and judicial oversight, sentencing by machine risks undermining constitutional legitimacy and eroding the moral foundations of justice