The study examines whether large language models truly base legal verdicts on the statutes or precedents they cite. By fixing case facts and swapping the cited authority with an unrelated one, researchers found that while models correctly name the authority 66.7%-100% of the time, the resulting verdict often remains unchanged—only 0.0%-21.7% of the time on CaseHOLD, 30.0%-76.7% on ECHR and SCOTUS, and 43.3%-50.0% on ContractNLI. Even a purpose‑built legal‑reasoning model and larger scales did not close this gap, and adversarial instructions further exposed vulnerabilities in verdict consistency.
By Saisab Sadhu, Shreeyans Arora, Pratinav Seth
arXiv:2608.21409v1 Announce Type: cross
Abstract: In medicine, claims remain valid when supported by empirical evidence grounded in stable biological reality. In law, by contrast, truth is contingent...
By Lorenzo Molfetta, Alessio Cocchieri, Luca Ragazzi, Ilaria Bartolini, Marco Patella, Gianluca Moro
arXiv:2604. 26233v3 Announce Type: replace Abstract: As Large Language Models (LLMs) are proposed as legal decision assistants, and even first-instance decision-makers, across a range of judicial and administrative contexts, it becomes essential to explore how they answer legal questions, and in particular the factors that lead them to decide difficult questions.
By Oisin Suttle, David Lillis
The study examines how Large Language Models (LLMs) can exhibit an ‘inertia of confidence’, giving incorrect legal verdicts with high certainty, and tests this on Indian Contract Act cases. Phase I audits ChatGPT, Meta AI, and Perplexity AI, introducing the High‑Confidence Error Rate (HCER) to measure dangerous certainty, finding Meta AI most prone to errors. Phase II surveys 380 Indian law students, revealing that exposure to hallucinated citations increases verification efforts but most students lack formal ethical AI training.
By Angel Mary John, Vipin Kumar Singh, Jerrin Thomas Panachakel
arXiv:2606. 23716v1 Announce Type: cross Abstract: Legal AI benchmark research frequently invokes the assumption that large language models can improve access to justice, including for people who cannot access lawyers in order to understand and exercise their legal rights.
By Andrew Lou, David Shin
The paper argues that hallucinations by legal language models should be judged as failures of legal warrant rather than mere factual or citation errors. It defines claim-authority warrant as a context-sensitive relationship between a legal claim and applicable, current authority, and proposes that evaluating warrant can uncover failures missed by traditional accuracy or citation metrics. The authors outline a pilot study, benchmark specifications, and a research agenda to assess whether legal AI systems’ claims are properly licensed by law.
By Maksym Taranukhin, Vered Shwartz
Legal Research Bench (LRB) is a new benchmark comprising 413 open-ended U.S. legal research questions, each paired with a gold answer, supporting authorities, and a binary grading rubric. The study evaluates thirteen advanced language‑model agents using web search, case‑law search, page parsing, and retrieval tools, scoring responses only when all required criteria are met and cited authorities verify. Results show that even the best model, Claude Opus 4.8, achieves full correctness on only 42.9% of questions, with performance varying by legal area and task complexity, and no clear link between more tool calls or inference cost and higher accuracy.
By Katrina Drozdov, Oliver Chen, Langston Nashold, Rayan Krishnan
The paper proposes a new method for evaluating AI accountability by analyzing the structural quality of a model’s defense for its decisions, using a four‑phase dialectical protocol based on Walton’s argumentation schemes and Govier’s criteria. Applied to nine large language models and 200 ambiguous moral-choice items, the study finds that models generally defend their reasoning well above the rubric minimum, though failures cluster on grounds and sufficiency and correlate with epistemic hedging. The protocol also reveals that models often present different argument schemes in justification than in reasoning, detects indefensible defenses, and highlights challenges in assessing retraction in AI alignment.
By Daan R. Henselmans, Derck W. E. Prinzhorn, Arno Libert
JusticeAxis is a new benchmark that formalizes legal judgment as a reference‑anchored task, requiring a single decision tied simultaneously to the statute and the surrounding circumstances. It comprises 256 real‑world criminal cases from 18 countries, each with audio, image, and text evidence, and three lawyer‑written judgments per case: the recorded judgment and two failure judgments. The authors also introduce JusticeAgent, a modular system that first establishes facts and then applies the law, with skills distilled from execution trajectories and validated under Bayesian credible bounds. Experiments reveal that open‑weight backbones tend to drift toward unsupported grounds while frontier models default to statutory interpretations, and that JusticeAgent can effectively plug into commercial systems.
By Zhengkai Tu, Mingda Zhang, Zijia Wang, Xiaoying Tang, Jimmy Huang
The study examines whether large language models (LLMs) can perform legally meaningful reasoning by testing OpenAI GPT 5.4 on European Court of Human Rights case forecasting. Using various prompting strategies, the authors find that the model produces structurally complete but substantively shallow analyses, and that LLM-as-a-Judge evaluators are internally consistent yet only weakly aligned with human annotators. The expert-curated prompt yields more comprehensive reasoning but does not improve prediction accuracy, leading the authors to caution against relying solely on automated LLM evaluation or using task accuracy as a proxy for reasoning quality.
By Amogh Raina, Ilias Chalkidis, Daniel Hershcovich, Henrik Palmer Olsen
The paper investigates whether large language model (LLM) chatbots can emulate human legal judgments of reasonableness. By comparing responses from 26 LLMs to those of human participants across 25 legal scenarios, the study finds that chatbots generally track human answers but tend to produce more homogeneous, government‑ and corporation‑friendly responses and align more closely with white, male, older, and more educated respondents. The authors note that these patterns warrant further systematic research.
By Nirav Patel, Emily Wenger, Christopher Buccafusco
The paper introduces a taxonomy of six user challenge types and a four-layer framework to analyze how large language models respond to user disagreement. Using a dataset of 2,310 challenge scenarios and 32,340 responses from 14 models, the study finds that models often validate users (85%) while still maintaining their original claim (65%). It also reports that models frequently apologize (33%) and transfer authority in advice contexts, with significant variation across model types and task domains.
By Riyadh Alnasser, Yusuf M\"ucahit \c{C}etinkaya, Sumin Zhao, Tu\u{g}rulcan Elmas