AI Diminishes the Condemnatory Character of Punishment: A Reply to Ryberg and Taylor

This article contributes to an ongoing debate in this journal regarding the impact of artificial intelligence (AI) on the condemnatory aspect of punishment, directly responding to Isaac Taylor’s argument that AI undermines condemnation in sentencing and Jesper Ryberg’s critique of this view. Rather than focusing on how algorithmic sentencing affects the condemnatory character of the sentencing process itself, as Taylor suggests, this article reframes the “expressive challenge” by examining AI sentencing tools’ potential impact on the condemnation inherent in the ultimate criminal punishment. Contrary to Taylor, it is argued that the sentencing process, understood as the method by which a judge arrives at a sentence, cannot inherently serve a condemnatory function. However, the manner in which a sentence is determined (e.g. through human deliberation versus an algorithm) can significantly affect the meaning of ultimate punishment. In response to Ryberg, the article suggests that the sentencing process can express other important attitudes, and ceding decision-making to AI may fail to express them properly. Ultimately, the article concludes that whether algorithmic sentencing negatively affects the condemnatory character of punishment is an empirical question, but it remains a concern that potential benefits must be measured against.