Metraipolis › archive › dc_934480daa056

Legal LLM Hallucination Should Be Evaluated as Failure of Legal Warrant — arXiv:2609.17546

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.17546
License (as archived)
CC0-1.0
Archived
2026-10-15T13:49:03.460Z
SHA-256 of the archived text
0f6546d722a8dbd353f2e6b6464553cc4d7dd7c8218df4e54a6fd937198e738c
Stable URL of this page
https://metraipolis.com/documento/dc_934480daa056

How to cite

Legal LLM Hallucination Should Be Evaluated as Failure of Legal Warrant — arXiv:2609.17546 — arXiv (en), https://arxiv.org/abs/2609.17546 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_934480daa056 (Metraipolis archive dc_934480daa056, sha256 0f6546d722a8dbd353f2e6b6464553cc4d7dd7c8218df4e54a6fd937198e738c)
BibTeX
@misc{dc934480daa056,
  title = {Legal LLM Hallucination Should Be Evaluated as Failure of Legal Warrant — arXiv:2609.17546},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.17546},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_934480daa056; sha256 0f6546d722a8dbd353f2e6b6464553cc4d7dd7c8218df4e54a6fd937198e738c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.17546. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

In this position paper, we argue that legal LLMs' hallucinations should be evaluated as a failure of legal warrant rather than as factual inaccuracy or citation failure. We define claim-authority warrant as the context-sensitive relation between a consequential legal claim and authority that exists, applies to the relevant jurisdiction, is current for the date of analysis, has the legal status represented by the system, and supports the proposition asserted. Warranted legal generation is the broader system behavior that answers, narrows, asks, warns, corrects a false premise, or abstains according to that relation. The falsifiable prediction is that warrant metrics reveal material failures that answer accuracy, citation existence, generic attribution, LegalHalBench-style statute relevance, and CitaLaw-style sentence-citation alignment can miss. We sharpen this claim with a side-by-side comparison item and a small, reproducible pilot over public-rule tests. We then specify benchmark records, claim boundaries, support labels, mixed response-policy scoring, risk weights, annotation reliability reporting, and jurisdiction-specific authority ontologies. The result is a concrete research agenda for evaluating legal AI systems by whether their consequential claims are licensed by law.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-17 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.17546 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-15T13:48:24.254Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-17?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).