Session: “Model Justice”

Session: “Model Justice”

After introducing the possibility of algorithms determining sentences in criminal cases – a scenario I call ‘model justice’ – and indicating some of its attractive features, this talk explores three objections to the idea involving (a) the role of the emotions in moral reasoning, (b) the particularism / generalism debate, and (c) the open question argument. The goal is to use these objections to highlight some desirable features of sentencing algorithms and, by extension, any algorithm designed to make morally significant decisions. I close with reflection on the relevance of these arguments to the alignment problem for artificial intelligence.