Stealthy LLM-Driven Data Poisoning Attacks Against Embedding-Based Retrieval-Augmented Recommender Systems

TL;DR

This paper investigates how small, subtle modifications to item descriptions in retrieval-augmented recommender systems can be exploited to unfairly promote or demote items, revealing significant vulnerabilities and the need for improved defenses.

Contribution

The study formalizes stealthy data poisoning attacks on RAG-based recommenders and demonstrates their effectiveness through experiments, highlighting critical security gaps.

Findings

Small token modifications can significantly alter item rankings.

Stealthy attacks evade naive detection methods.

Robust textual consistency checks are necessary for defense.

Abstract

We present a systematic study of provider-side data poisoning in retrieval-augmented recommender systems (RAG-based). By modifying only a small fraction of tokens within item descriptions -- for instance, adding emotional keywords or borrowing phrases from semantically related items -- an attacker can significantly promote or demote targeted items. We formalize these attacks under token-edit and semantic-similarity constraints, and we examine their effectiveness in both promotion (long-tail items) and demotion (short-head items) scenarios. Our experiments on MovieLens, using two large language model (LLM) retrieval modules, show that even subtle attacks shift final rankings and item exposures while eluding naive detection. The results underscore the vulnerability of RAG-based pipelines to small-scale metadata rewrites and emphasize the need for robust textual consistency checks and provenance tracking to thwart stealthy provider-side poisoning.