Who Checks the Citations? Benchmarking Legal Hallucination Detection

• URL: http://arxiv.org/abs/2606.21155v1
• Date: Fri, 19 Jun 2026 06:47:31 GMT
• Title: Who Checks the Citations? Benchmarking Legal Hallucination Detection
• Abstract summary: Attorneys, judges, and pro se filers increasingly use AI to draft legal documents, yet these tools frequently fabricate citations.<n>This study evaluates whether AI-based systems can mitigate these errors by automatically detecting hallucinations.<n>We propose a taxonomy of legal citation hallucinations grounded in actual court filings and introduce a dataset of 1,300 brief excerpts containing injected errors.
• Score: 7.592193069681834
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Attorneys, judges, and pro se filers increasingly use AI to draft legal documents, yet these tools frequently fabricate citations. Despite predictions that newer models would hallucinate less or that court sanctions would deter negligent filers, we found over 1,000 filings containing fabricated citations -- with this number growing year-over-year. This study evaluates whether AI-based systems can mitigate these errors by automatically detecting hallucinations. We propose a taxonomy of legal citation hallucinations grounded in actual court filings and introduce a dataset of 1,300 brief excerpts containing injected errors. Benchmarking five models in agentic and non-agentic settings reveals that while the latest iterations perform better -- GPT-5 achieves 82.8% recall and a 60.5% F1 score in an agentic framework -- all models struggle with subtle error categories. Agentic verification remains resource-intensive, with GPT-5 averaging 16.9 steps per excerpt. Furthermore, restricted information access limits the efficacy of even the best agents. This gap creates policy concerns, as it disadvantages both AI systems and litigants who lack subscriptions to commercial legal databases. Together, our dataset, tools, and policy recommendations provide a foundation for building and auditing reliable legal citation checking tools.

Related papers

• NovGauge: A Fine-Grained Benchmark for Diagnosing LLMs' Capability in Paper Novelty Assessment [54.4265627247104]
  Large language models (LLMs) are increasingly used in peer review at major AI conferences.<n>Existing benchmarks assess novelty as a single holistic score.<n>We present NovGauge, a human-anchored benchmark for fine-grained novelty assessment diagnosis.
  arXiv  Detail & Related papers  (2026-09-10T08:34:56Z)
• Who is the Agent to Blame? Localizing Faithfulness and Citation Mistakes in Agentic Deep Research [75.8770212601511]
  Deep research (DR) systems produce long-form cited reports by orchestrating multiple agents that search and synthesize information from the web.<n> Citations are the primary mechanism for evaluating the faithfulness of these reports, yet current DR systems exhibit poor citation recall.<n>We propose an evaluation method that pinpoints which agent introduced each error by locally testing agent invocations for faithfulness and verifiability relative to their own inputs.
  arXiv  Detail & Related papers  (2026-08-25T09:34:18Z)
• From Process to Evidence: How Computing Can Ground Appropriate Reliance on Legal AI [0.0]
  After more than 1,500 cases involving AI hallucinations, lawyers have been instructed to perform careful, independent review of AI-assisted filings.<n>Discharging these duties requires what the human-computer interaction literature calls appropriate reliance''<n>We analyze the official record of the New York court system.
  arXiv  Detail & Related papers  (2026-07-30T22:24:59Z)
• Fantastic Adaptive Taxonomies and How to Use Them [100.20614173157708]
  AdaMAST converts execution traces into a compact, evidence-grounded failure taxonomy.<n>No code is hand-authored, no trace human-annotated.<n>In agent-system search, taxonomy-coded diagnoses of failed candidates outperform free-form.
  arXiv  Detail & Related papers  (2026-07-17T17:41:16Z)
• SynCred-Bench: Benchmarking Synthetic Credibility in AI-Generated Visual Misinformation [82.22252244442744]
  SYNCRED-Bench is a benchmark of 600 AI-generated misinformation images balanced across six credible-form categories and seven fine-grained circulation styles.<n>Under a 5% false-positive-rate constraint, 15 MLLMs achieve only 10.5% true positive rate (TPR), open-source AIGC detectors achieve less than 5%.<n>Human annotators also struggled to identify synthetic credibility, reaching only 63% TPR.
  arXiv  Detail & Related papers  (2026-06-02T08:57:38Z)
• Citation Grounding: Detecting and Reducing LLM Citation Hallucinations via Legal Citation Graphs [0.0]
  Empirical evaluation on 100 Ukrainian legal queries across five systems.<n>13-21% of citations hallucinated, with citation precision ranging from 0.791 to 0.873.<n>To reduce hallucinations without human annotation, we introduce Citation Grounding DPO.
  arXiv  Detail & Related papers  (2026-05-30T21:22:47Z)
• HalluCiteChecker: A Lightweight Toolkit for Hallucinated Citation Detection and Verification in the Era of AI Scientists [58.87954687016989]
  HalluCiteChecker is a toolkit for detecting and verifying hallucinated citations in scientific papers.<n>Our code is released under the Apache 2.0 license on GitHub and is distributed as an installable package via PyPI.
  arXiv  Detail & Related papers  (2026-04-29T16:01:42Z)
• CiteAudit: You Cited It, But Did You Read It? A Benchmark for Verifying Scientific References in the LLM Era [51.63024682584688]
  Large language models (LLMs) introduce a new risk: fabricated references that appear plausible but correspond to no real publications.<n>We present the first comprehensive benchmark and detection framework for hallucinated citations in scientific writing.<n>Our framework significantly outperforms prior methods in both accuracy and interpretability.
  arXiv  Detail & Related papers  (2026-02-26T19:17:39Z)
• Compound Deception in Elite Peer Review: A Failure Mode Taxonomy of 100 Fabricated Citations at NeurIPS 2025 [0.0]
  Large language models (LLMs) are increasingly used in academic writing, yet they frequently hallucinate by generating citations to sources that do not exist.<n>This study analyzes 100 AI-generated hallucinated citations that appeared in papers accepted by the 2025 Conference on Neural Information Processing Systems.<n>Despite review by 3-5 expert researchers per paper, these fabricated citations evaded detection, appearing in 53 published papers.
  arXiv  Detail & Related papers  (2026-02-05T17:43:35Z)
• The 17% Gap: Quantifying Epistemic Decay in AI-Assisted Survey Papers [0.0]
  "Hallucinated papers" are a known artifact, but the systematic degradation of valid citation chains remains unquantified.<n>We conducted a forensic audit of 50 recent survey papers in Artificial Intelligence published between September 2024 and January 2026.<n>We detect a persistent 17.0% Phantom Rate -- citations that cannot be resolved to any digital object despite aggressive forensic recovery.
  arXiv  Detail & Related papers  (2026-01-24T12:00:55Z)
• HalluGraph: Auditable Hallucination Detection for Legal RAG Systems via Knowledge Graph Alignment [0.0]
  We introduce HalluGraph, a graph-theoretic framework that quantifies hallucinations through structural alignment between knowledge graphs extracted from context, query, and response.<n>Our approach produces bounded, interpretable metrics into textitEntity Grounding (EG), measuring whether entities in the response appear in source documents, and textitRelation Preservation (RP), verifying that asserted relationships are supported by context.
  arXiv  Detail & Related papers  (2025-12-01T13:31:06Z)
• Security Challenges in AI Agent Deployment: Insights from a Large Scale Public Competition [101.86739402748995]
  We run the largest public red-teaming competition to date, targeting 22 frontier AI agents across 44 realistic deployment scenarios.<n>We build the Agent Red Teaming benchmark and evaluate it across 19 state-of-the-art models.<n>Our findings highlight critical and persistent vulnerabilities in today's AI agents.
  arXiv  Detail & Related papers  (2025-07-28T05:13:04Z)
• CLERC: A Dataset for Legal Case Retrieval and Retrieval-Augmented Analysis Generation [44.67578050648625]
  We transform a large open-source legal corpus into a dataset supporting information retrieval (IR) and retrieval-augmented generation (RAG) This dataset CLERC is constructed for training and evaluating models on their ability to (1) find corresponding citations for a given piece of legal analysis and to (2) compile the text of these citations into a cogent analysis that supports a reasoning goal.
  arXiv  Detail & Related papers  (2024-06-24T23:57:57Z)
• Hallucination-Free? Assessing the Reliability of Leading AI Legal Research Tools [32.78336381381673]
  We report on the first preregistered empirical evaluation of AI-driven legal research tools. We find that the AI research tools made by LexisNexis (Lexis+ AI) and Thomson Reuters (Westlaw AI-Assisted Research and Ask Practical Law AI) each hallucinate between 17% and 33% of the time. It provides evidence to inform the responsibilities of legal professionals in supervising and verifying AI outputs.
  arXiv  Detail & Related papers  (2024-05-30T17:56:05Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.