AI/ legal ai · hallucination · benchmarks · retrieval

Fine-Tuned Legal AI Hallucinates More When Statutes Are Missing

A benchmark built on fire-safety regulations finds that domain-adapted models confabulate more than generalists when key statutory context is absent.

Fine-tuning a language model on legal statutes makes it more likely to confabulate when the relevant law is missing, not less.

Researchers published SearchFireSafety, a benchmark designed to test how well LLMs handle statute-based legal questions, a domain that existing evaluations mostly ignore in favor of case law. Unlike judicial opinions, which read as self-contained narratives, statutes are organized in hierarchical trees: a provision may only make sense alongside parent sections and cross-referenced definitions scattered across an entire regulatory code. The benchmark uses fire-safety regulations as its test case, pairing real questions that require citation-aware retrieval with synthetic scenarios where critical statutory context is deliberately withheld. Graph-guided retrieval, which treats the statutory hierarchy as a connected graph rather than a flat document set, substantially improved accuracy across the LLMs tested.

The core finding cuts against the usual logic of specialization: models adapted to the legal domain were more likely to fabricate answers when supporting statutes were absent. In regulatory settings where accuracy is non-negotiable, a model that confidently cites an inapplicable rule is more dangerous than one that simply refuses to answer.

Most hallucination benchmarks, and most legal-AI evaluations built on case law, may give vendors a false sense of safety for exactly the statutory use cases they are actively selling.

TR

The Revision

Written by an AI system from the public sources credited above. How we write →