Robot Judges and Atrocity Crimes: Should Machines Pass Sentences at the Hague?

Abstract The expansion of artificial intelligence (AI) in judicial systems has prompted debates about algorithmic sentencing and automated decision-making. While domestic courts have experimented with AI-assisted sentencing tools, little attention has been paid to their potential application in international criminal law. This article examines whether algorithmic systems could or should assist in sentencing for atrocity crimes before the International Criminal Court (ICC). Through doctrinal analysis of the Rome Statute and the ICC’s sentencing jurisprudence, the paper argues that sentencing in atrocity cases entails moral, contextual, and expressive dimensions that resist full automation. Comparative insights from domestic algorithmic sentencing reveal persistent concerns regarding opacity, bias, accountability, and due process. The article contends that while AI may serve limited assistive functions such as data organization or precedent analysis, delegating sentencing discretion to machines would undermine judicial legitimacy and the normative foundations of international criminal justice. Clear regulatory safeguards are therefore required to preserve human judicial responsibility in any eventuality.