Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Incorporating Legal Structure in Retrieval-Augmented Generation: A Case Study on Copyright Fair Use

Justin Ho, Alexandra Colby, William Fisher

TL;DR

This work tackles the problem of unreliable, text-only retrieval in legal AI applied to the Fair Use Doctrine under DMCA takedown contexts. It introduces a structured Retrieval-Augmented Generation framework that combines semantic search with a knowledge graph of legal precedents, court hierarchies, and factor-level content, augmented by citation networks and Chain-of-Thought reasoning. A functioning Neo4j-based prototype demonstrates that incorporating legal structure improves doctrinal relevance (as measured by PageRank-derived authority) at the expense of raw textual similarity, highlighting a trade-off between law-focused grounding and surface similarity. While promising, the approach requires rigorous evaluation, time-aware authority measures, and extension to other legal doctrines to enhance accessibility and reliability of automated legal assistance.

Abstract

This paper presents a domain-specific implementation of Retrieval-Augmented Generation (RAG) tailored to the Fair Use Doctrine in U.S. copyright law. Motivated by the increasing prevalence of DMCA takedowns and the lack of accessible legal support for content creators, we propose a structured approach that combines semantic search with legal knowledge graphs and court citation networks to improve retrieval quality and reasoning reliability. Our prototype models legal precedents at the statutory factor level (e.g., purpose, nature, amount, market effect) and incorporates citation-weighted graph representations to prioritize doctrinally authoritative sources. We use Chain-of-Thought reasoning and interleaved retrieval steps to better emulate legal reasoning. Preliminary testing suggests this method improves doctrinal relevance in the retrieval process, laying groundwork for future evaluation and deployment of LLM-based legal assistance tools.

Incorporating Legal Structure in Retrieval-Augmented Generation: A Case Study on Copyright Fair Use

TL;DR

This work tackles the problem of unreliable, text-only retrieval in legal AI applied to the Fair Use Doctrine under DMCA takedown contexts. It introduces a structured Retrieval-Augmented Generation framework that combines semantic search with a knowledge graph of legal precedents, court hierarchies, and factor-level content, augmented by citation networks and Chain-of-Thought reasoning. A functioning Neo4j-based prototype demonstrates that incorporating legal structure improves doctrinal relevance (as measured by PageRank-derived authority) at the expense of raw textual similarity, highlighting a trade-off between law-focused grounding and surface similarity. While promising, the approach requires rigorous evaluation, time-aware authority measures, and extension to other legal doctrines to enhance accessibility and reliability of automated legal assistance.

Abstract

This paper presents a domain-specific implementation of Retrieval-Augmented Generation (RAG) tailored to the Fair Use Doctrine in U.S. copyright law. Motivated by the increasing prevalence of DMCA takedowns and the lack of accessible legal support for content creators, we propose a structured approach that combines semantic search with legal knowledge graphs and court citation networks to improve retrieval quality and reasoning reliability. Our prototype models legal precedents at the statutory factor level (e.g., purpose, nature, amount, market effect) and incorporates citation-weighted graph representations to prioritize doctrinally authoritative sources. We use Chain-of-Thought reasoning and interleaved retrieval steps to better emulate legal reasoning. Preliminary testing suggests this method improves doctrinal relevance in the retrieval process, laying groundwork for future evaluation and deployment of LLM-based legal assistance tools.
Paper Structure (17 sections, 1 equation, 5 figures, 2 tables)

This paper contains 17 sections, 1 equation, 5 figures, 2 tables.

Table of Contents

  1. Introduction
  2. LLMs in Legal Assistance
  3. Local Expertise and Structured Reasoning
  4. Literature Review and Related Work
  5. Similarity is Not All You Need
  6. Incorporating Legal Structure
  7. Methods
  8. Data Corpus
  9. Data Representation
  10. Retrieval and Reranking
  11. Current Progress and Implementation
  12. Preliminary Testing on Retrieval
  13. Limitations and Future Work
  14. Future Evaluation
  15. Limitations of Current Work
  16. ...and 2 more sections

Figures (5)

  • Figure 1: Overview of the Automated Analysis of Fair Use Cases.
  • Figure 2: Domain Specific Knowledge Graph Schema of Judicial Opinions of the Fair Use Doctrine in Copyright 17 U.S. Code § 107.
  • Figure 3: Distribution of Legal Case Influence by Court and PageRank
  • Figure 4: Interface of the Fair Use Legal Bot
  • Figure 5: Comparison of Retrieval Methods. The Standard RAG achieves higher textual similarity but lacks doctrinal authority as measured by PageRank, and the most common cases retrieved tend to be less authoritative.