ClaimTrust introduces a dynamic framework for evaluating document reliability in retrieval-augmented generation systems, addressing critical security concerns in AI-generated content.

• Implements a modified PageRank algorithm to propagate trust scores across document relationships
• Dynamically evaluates reliability of information sources used in RAG pipelines
• Creates a security layer to detect and filter potential fake news and misinformation
• Enhances overall trustworthiness of AI-generated content for business applications

This research is vital for security professionals as it provides a systematic approach to mitigate misinformation risks in enterprise AI deployments, protecting organizations from reputational damage and security breaches caused by untrustworthy information sources.

ClaimTrust: Propagation Trust Scoring for RAG Systems