Sotomayor says AI forecasting Supreme Court decisions a 'bad thing,' shows 'we're way too predictable'

Supreme Court Justice Sonia Sotomayor declared on Thursday, March 27, 2025, during an event at the University of Alabama School of Law in Tuscaloosa, that the growing accuracy of artificial intelligence models in predicting the Court's rulings is a "very bad thing," expressing concern that it reveals the judiciary is becoming overly formulaic and risks undermining the nuanced, case-by-case analysis fundamental to justice. Her remarks were made in response to recent academic studies and commercial tools that have demonstrated a high success rate in forecasting Supreme Court decisions by analyzing patterns in legal language, precedent, and the justices' past rulings. Justice Sotomayor elaborated that predictability, while valued for legal stability, can be detrimental when it suggests the Court is merely applying a rigid algorithm rather than engaging in the deep, contextual deliberation required for each unique case. She warned that if AI can reliably map the Court's logic, it might indicate the justices are not sufficiently challenging their own assumptions or evolving with societal changes. This, she argued, could erode public confidence in the Court as an independent branch capable of surprising, yet reasoned, judgments that adapt constitutional principles to new realities. The broader context of her comments touches on a significant debate within the legal community about the role of technology in law. While some scholars see predictive AI as a tool for efficiency and access to justice, others, like Sotomayor, view it as a mirror reflecting potential systemic flaws. Her critique is not aimed at the technology itself but at what its success implies about judicial behavior. It serves as a cautionary note for her colleagues and the legal profession to guard against falling into predictable patterns that could be gamed by litigants or perceived as mechanical, thereby threatening the Court's legitimacy and its role as a dynamic interpreter of the law.