Doron Narotzki (Akron) has a new working paper on SSRN, Predicting Tariff Decisions from Oral Argument Transcripts with AI Models. Here is the abstract:

This Article uses a single Supreme Court tariffs case as a case study to explore how large language models (LLMs) interpret oral argument and predict case outcomes. The goal is to observe how different LLMs process the same argument transcript, how they translate questioning patterns into probabilistic forecasts, and what these outputs reveal about the models’ implicit picture of judicial reasoning. By limiting the AI models to the argument transcript only, this project seeks to explore how different LLMs would predict the Supreme Court’s decision on a legally and economically significant tariffs dispute and will later assess how those predictions align with the opinion the Court ultimately issues.