'Transparency versus explanation: The role of ambiguity in legal AI' by Elena Esposito

Transparency versus explanation: The role of ambiguity in legal AI

In our sixth online-first article (November 2021), Elena Esposito argues that the the goal of ML explanation is not to reveal the procedures of such systems, but to communicate with them and to obtain relevant and controlled information that is relevant to external (legal) systems. Legal arguments require interpretability, implying not that artificial legal intelligence systems provide insufficient explanation, but that any explanation they give is likely to be too precise -- thus constraining the freedom of interpretation and contestability that law requires. 

Reply by Federico Cabitza, University of Milan-Bicocca.