Where Experts Disagree, Models Fail: Detecting Implicit Legal Citations in French Court Decisions
arXiv:2603.22973
The paper introduces a benchmark for detecting implicit citations of the French Civil Code in court decisions and shows that cases where legal experts disagree are especially hard for AI models, with an ensemble achieving an overall F1 of 0.70.
Abstract
Applying computational methods to law at scale requires separating genuine legal reasoning from surface similarity. We study this through a concrete task: detecting implicit citations of the French Civil Code, where a court applies a statutory rule without naming it: a post-hoc question about the reasoning a court actually used. We release a benchmark of 1,015 passage-article pairs annotated by three legal experts. Our central finding is that their disagreement is itself informative: the third of cases the experts dispute are where models fail. Our best ensemble reaches an F1 score of 0.70 overall. Yet, two-thirds of its false positives fall on those disputed cases, a concentration that holds across all ten models we evaluate. Disagreement is a signal of intrinsic difficulty, not annotation noise. This should not block useful tools, however: reframed as top- ranking with multi-model consensus, the same signals reach 76% precision for the top-200 candidates without supervision.