Artificial Testimony: Algorithms and Legal Proof

When a person makes a decision, and that decision results in harm, often that person can be called in to court to explain their actions. What happens when an algorithmic decision results in harm? One option is to create new legal structures to govern algorithms; but another option is to create algorithms capable of meeting existing legal standards of proof. This talk presents three modes by which algorithmic can be understood – transparency, statistical analysis, and explanation – and describes how each could fit into the Anglo-American legal tradition.

Ce contenu a été mis à jour le 1 septembre 2020 à 14 h 20 min.