Social sciences · OECD FoS 5.5 · China MoE 0301 法学
Law
法学
L1Assistive ↑ Rising Score 1.00 · Last reviewed 2026-10-21 · Next review 2027-01-19
BIGGEST BREAKTHROUGH 2026-10-18 AI · incrementalField · major
Appeals court upholds Thomson Reuters’ win over Ross Intelligence on AI-training fair use (opinion sealed) ↗
AI: fair-use defence for training data fails on appeal; Field: an appellate-level ruling on AI training copyright (details pending)
Why this level
Legal practice uses AI widely, with problems: Charlotin's database lists 2,202 decisions on AI hallucinations, Reuters counts 1,395 US cases; >60% of surveyed federal judges have used AI but only 22.4% weekly.
S = 0.30·method + 0.35·discovery + 0.25·adoption + 0.10·education (education scores 0 when there is no evidence; if empty, the other weights are renormalised)
Sources
- damiencharlotin.com/hallucinations
- reuters.com/legal/government/ai-error-ridden-court-filings-surge-despite-three-years-court-sanctions-2026-09-17
- nycbar.org/reports/artificial-intelligence-in-federal-courts-a-random-sample-survey-of-judges
Composite 0.75–1.49: AI is an aid in parts of research or practice, with cases or surveys, but no recognised AI-key result. All levels →
Research directions
Legal research & draftingContract reviewJudicial assistanceAI law (as subject)
Papers
- Hallucination‐Free? Assessing the Reliability of Leading
AI
Legal Research Tools
First independent test of commercial legal AI tools
2025-04-23 · Journal of Empirical Legal Studies
- Large Legal Fictions: Profiling Legal Hallucinations in Large Language Models
Quantifies LLM hallucination on case-law questions
2024-06-26 · Journal of Legal Analysis
- GPT-4 passes the bar exam
GPT-4 passes the Uniform Bar Exam
2024-02-26 · Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences
- Commit or Abstain? System One Decision Models, LLMs, and Missing Legal Authority
Tests whether calibrated “System One” decision models abstain when the correct legal authority is missing.
2026-10-21 · arXiv · M. Mikail Demir
- An LLM-Based Agent for Legal Compliance Assessment: Architecture, Memory, and Evaluation in a Canadian and Quebec Context
An LLM compliance agent grounded in a legal corpus, evaluated on 13 Canadian and Quebec instruments.
2026-10-19 · arXiv · Soukaina Alaoui Ismaili, Jean-Sebastien Dessureault
- StrucChunk: Structure-Aware Retrieval for Legal Document RAG
Structure-aware chunking and hybrid retrieval for hierarchical legal documents.
2026-10-19 · arXiv · Surya Deep Singh, Ashutosh Tripathi, Gaurav Deep Singh
- LLMs Prompted for Legal Context Refuse More: Overrefusal from Small On-Premises LLMs in Criminal Legal Context
Stating a lawyer role makes small on-prem LLMs over-refuse benign criminal-law requests.
2026-10-19 · arXiv · Anastasiia Kucherenko, Francois Brouchoud, Dimitri Percia David, Andrei Kucharavy
- Bilingual Legal Citation Fabrication Benchmark: Saudi PDPL and GDPR (Arabic–English)
A 120-question Arabic–English benchmark testing whether LLMs cite correct GDPR/PDPL articles or fabricate.
2026-10-19 · OpenAlex · arXiv (Cornell University) · Noura Alrajeh
- Corporate Accountability by Algorithm: AI and the Interpretation of Human Rights Due Diligence
Examines AI tools for human-rights due diligence and their risks to corporate accountability.
2026-10-19 · OpenAlex · Business and Human Rights Journal · Galit A. Sarfaty
- A Response to “How AI Destroys Institutions”
A law-review response to the argument that AI systems inherently destroy institutions.
2026-10-19 · OpenAlex · SMU Law Review Forum · Andrew M. Perlman
- Evidence-Grounded LLM-Assisted Android Mobile Forensic Reporting
Claim-level evaluation of LLM-drafted Android forensic reports reveals large differences in citation validity.
2026-10-18 · arXiv · Madhab Chandra Das, Sajid Ahmed, Tasnia Tabassum, Qingzhong Liu
- From arXiv, bioRxiv, medRxiv and OpenAlex (last ~30 days); summaries written by AIMap from the abstract.
Milestones 7
Big glowing nodes are leaps: AI leap cool · Field leap warm · double ring = both
-
Legal
Appeals court upholds Thomson Reuters’ win over Ross Intelligence on AI-training fair use (opinion sealed)AI: fair-use defence for training data fails on appeal; Field: an appellate-level ruling on AI training copyright (details pending)
The VergeTo be verified -
Industry
Reuters: AI error-ridden court filings surge despite sanctionsAI-error filings keep surging
Reuters -
Industry
NYC Bar random-sample survey: 60%+ of federal judges have used AI, ~22% daily or weeklyFirst random-sample survey of judges’ AI use
New York City Bar -
Paper
Hallucination-free? Assessing AI legal research tools (JELS)First independent test of commercial legal AI tools
Journal of Empirical Legal Studies -
Paper
Large Legal Fictions: profiling legal hallucinations in LLMs (J. Legal Analysis)Quantifies LLM hallucination on case-law questions
Journal of Legal Analysis -
Competition
GPT-4 passes the bar exam (Phil. Trans. R. Soc. A)GPT-4 passes the Uniform Bar Exam
Phil. Trans. R. Soc. A -
Regulatory
Mata v. Avianca: SDNY sanctions lawyers for six ChatGPT-fabricated casesLandmark sanction that triggered court AI rules
Reuters
No milestones match this filter.
More papers & repos 3
| Type | Title | Date |
|---|---|---|
| Paper | Hallucination‐Free? Assessing the Reliability of Leading AI Legal Research Tools | 2025-04-23 |
| Paper | Large Legal Fictions: Profiling Legal Hallucinations in Large Language Models | 2024-06-26 |
| Paper | GPT-4 passes the bar exam | 2024-02-26 |
Latest news 12
- What Happens When Lawyers Can Afford to Look Further?Artificial Lawyeren
- More Legal Research Moves as Moonlit Launches Free CommentariesArtificial Lawyeren
- Big Law’s Greatest GiftsArtificial Lawyeren
- Harvey Webinar – Agents in Legal: A Practical GuideArtificial Lawyeren
- Artificial Intelligence in Federal Courts: A Random Sample Survey of JudgesNew York City Baren