ARBITER: AI-Driven Filtering for Role-Based Access Control

2025-12-23

AI Summary

In dynamic enterprise environments, conventional Role-Based Access Control (RBAC) fails to prevent sensitive information leakage in Retrieval-Augmented Generation (RAG) systems due to LLM-induced issues—including prompt truncation, misclassification, and context loss—thereby compromising fine-grained document access control. Method: We propose an AI-driven, fine-grained dynamic access control framework that deeply integrates RBAC into the entire RAG pipeline—without classifier fine-tuning. Our approach leverages few-shot LLMs with prompt engineering for instantaneous role-policy deployment and agile updates; incorporates role-aware retrieval, hierarchical input/output validation, and post-generation factual verification to establish an end-to-end security loop. Contribution/Results: Evaluated on a synthetic dataset of 389 queries, our framework achieves 85% accuracy and 89% F1-score—matching traditional RBAC performance—and demonstrates, for the first time, the practical feasibility of dynamic access control natively within LLM-powered RAG systems.

Technology Category

• Natural Language Processing: Safety and Robustness
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Role-Based Access Control (RBAC) struggles to adapt to dynamic enterprise environments with documents that contain information that cannot be disclosed to specific user groups. As these documents are used by LLM-driven systems (e.g., in RAG) the problem is exacerbated as LLMs can leak sensitive data due to prompt truncation, classification errors, or loss of system context. We introduce our, a system designed to provide RBAC in RAG systems. our implements layered input/output validation, role-aware retrieval, and post-generation fact-checking. Unlike traditional RBAC approaches that rely on fine-tuned classifiers, our uses LLMs operating in few-shot settings with prompt-based steering for rapid deployment and role updates. We evaluate the approach on 389 queries using a synthetic dataset. Experimental results show 85% accuracy and 89% F1-score in query filtering, close to traditional RBAC solutions. Results suggest that practical RBAC deployment on RAG systems is approaching the maturity level needed for dynamic enterprise environments.

Problem

• Adapt RBAC to dynamic enterprise environments with sensitive documents
• Prevent LLM data leakage in RAG systems via access control
• Implement flexible, prompt-based RBAC for rapid deployment and updates

Innovation

• Layered input/output validation for RBAC
• Role-aware retrieval in RAG systems
• Post-generation fact-checking using LLMs