Anthony Niblett (U Toronto Law) and Albert Yoon (same) have posted “A.I. and the Nature of Disagreement” on SSRN. Here is the abstract:
Some legal commentators – including ourselves – have been loudly optimistic about the power of artificial intelligence (AI) to improve litigation. These commentators argue that AI can provide clearer information, cutting through much of the complexity of the law, reducing frictions and disagreements between the parties. Further, the possibility of using AI to determine the outcomes of legal disputes has given rise to the concept of “robot judges” in legal scholarship.
But in this paper, we argue that much of this literature fails to fully appreciate what litigated disputes are really about. Litigants may disagree about the facts of the case, the applicable rules, or how the rules apply to the facts. These disagreements are often complex and intertwined.
We contend that AI tools may be limited in their ability to resolve litigated disputes because these tools often address only one type of disagreement, leaving others unresolved. The optimistic view of AI in litigation assumes that parties disagree mainly about the likelihood of winning or the size of damages awards for a given set of agreed facts. But we question whether litigation is really fueled by such disagreements.
Our main takeaway is that if litigation is driven by disagreements over the facts or which rules should govern, AI’s capacity to reduce disagreement may fall short of what some proponents of AI claim. We call for more empirical and theoretical work to explore what litigants actually disagree about to better assess the likely impact of algorithmic decision-making in legal systems.
