Who Stole Your Data? A Method for Detecting Unauthorized RAG Theft

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Liu, Peiyang, Cui, Ziqiang, Liang, Di, Ye, Wei
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914082392637440
author Liu, Peiyang
Cui, Ziqiang
Liang, Di
Ye, Wei
author_facet Liu, Peiyang
Cui, Ziqiang
Liang, Di
Ye, Wei
contents Retrieval-augmented generation (RAG) enhances Large Language Models (LLMs) by mitigating hallucinations and outdated information issues, yet simultaneously facilitates unauthorized data appropriation at scale. This paper addresses this challenge through two key contributions. First, we introduce RPD, a novel dataset specifically designed for RAG plagiarism detection that encompasses diverse professional domains and writing styles, overcoming limitations in existing resources. Second, we develop a dual-layered watermarking system that embeds protection at both semantic and lexical levels, complemented by an interrogator-detective framework that employs statistical hypothesis testing on accumulated evidence. Extensive experimentation demonstrates our approach's effectiveness across varying query volumes, defense prompts, and retrieval parameters, while maintaining resilience against adversarial evasion techniques. This work establishes a foundational framework for intellectual property protection in retrieval-augmented AI systems.
format Preprint
id arxiv_https___arxiv_org_abs_2510_07728
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Who Stole Your Data? A Method for Detecting Unauthorized RAG Theft
Liu, Peiyang
Cui, Ziqiang
Liang, Di
Ye, Wei
Information Retrieval
Computation and Language
Retrieval-augmented generation (RAG) enhances Large Language Models (LLMs) by mitigating hallucinations and outdated information issues, yet simultaneously facilitates unauthorized data appropriation at scale. This paper addresses this challenge through two key contributions. First, we introduce RPD, a novel dataset specifically designed for RAG plagiarism detection that encompasses diverse professional domains and writing styles, overcoming limitations in existing resources. Second, we develop a dual-layered watermarking system that embeds protection at both semantic and lexical levels, complemented by an interrogator-detective framework that employs statistical hypothesis testing on accumulated evidence. Extensive experimentation demonstrates our approach's effectiveness across varying query volumes, defense prompts, and retrieval parameters, while maintaining resilience against adversarial evasion techniques. This work establishes a foundational framework for intellectual property protection in retrieval-augmented AI systems.
title Who Stole Your Data? A Method for Detecting Unauthorized RAG Theft
topic Information Retrieval
Computation and Language
url https://arxiv.org/abs/2510.07728