pipette
ESEspañol

From One Query to Many: A Critical Survey of Privacy Accounting in RAG

Zhiyuan Zhang

Preprint

In the authors' words

Retrieval-augmented generation (RAG) enables large language models to leverage external knowledge but introduces privacy risks when the retrieved corpus contains sensitive information. Differential privacy (DP) offers formal protection, yet interpreting and comparing privacy guarantees across existing DP-RAG methods remains challenging. In particular, a privacy budget reported for a single response does not directly characterize privacy loss over repeated queries, and even identical values of may represent different guarantees depending on the protected data unit, accounting horizon, and observable outputs. This survey critically examines how privacy accounting evolves from single-query generation to multi-query RAG systems. We categorize existing approaches into query-time privatization, multi-query accounting, and one-time corpus privatization. We analyze their underlying assumptions and privacy--utility trade-offs. Rather than ranking methods by alone, we propose a framework that interprets a DP-RAG guarantee along three dimensions: privacy granularity, accounting scope, and release boundaries. We conclude that an is a claim only together with the three slots: the protected unit, the accounting scope, and the release boundary.

Main resultThe abstract does not state a limitation.

Appeared: Friday, October 16. arXiv. Preprint, not yet peer-reviewed.