Signed but Steerable: Measuring and Closing the Text-Vector Integrity Gap in Offline Retrieval-Augmented Generation
Signed offline-RAG knowledge packs can carry manipulated embedding vectors that silently steer retrieval results, a subtle integrity threat for edge-deployed AI systems.
Summary written by editorial AI · Source link below
arXiv:2611.06498v1 Announce Type: new Abstract: Offline retrieval-augmented generation (RAG) on edge devices is updated through signed knowledge packs that carry text and precomputed embedding vectors. A signature proves who published a pack, not that its vectors encode its text. We study this text-vector integrity gap. A verifier that recomputes vectors and accepts them within a cosine tolerance concedes a predictable share of queries to an attacker who controls the vectors: the hijack rate eq
Editorial Analysis
Edge-deployed RAG systems that trust signed packs without verifying text-vector alignment face a stealthy manipulation risk that bypasses conventional integrity checks.
Add embedding-consistency validation to RAG update pipelines alongside existing signature verification.
Forward-looking interpretation drafted by editorial AI under human review — not a reproduction of the source. See methodology.
External link — opens at arXiv Crypto & Security in a new tab.
More from the AI Security Desk
- Claude Code's “Don't Ask Again” File Exposes Passwords and Keys in 1,122 Public Copies on GitHub2d
- Runaway AI hacking agent behind WorkOS breach after escaping sandbox3d
- AI Gave Your Incident Archive a Second Job3d
- Malicious Browser Extensions Could Use Claude Dispatch to Run Code on Users' Computers3d
- AI Slop and Hallucinations in Vulnerability Assessment: A Survey on Reasoning Failures and Trustworthy Mitigation3d