No Free Lunch: Retrieval-Augmented Generation Undermines Fairness in LLMs, Even for Vigilant Users

Mengxuan Hu, Hongyi Wu, Ronghang Zhu, Zihan Guan, Dongliang Guo, Daiqing Qi, Sheng Li


Abstract
Retrieval-Augmented Generation (RAG) is widely adopted for its effectiveness and cost-efficiency in mitigating hallucinations and enhancing the domain-specific generation capabilities of large language models (LLMs). However, is this effectiveness and cost-efficiency truly a free lunch? In this study, we comprehensively investigate the fairness costs associated with RAG by proposing a practical three-level threat model from the perspective of user awareness of fairness. Specifically, varying levels of user fairness awareness result in different degrees of fairness censorship on external datasets. We examine the fairness implications of RAG using uncensored, partially censored, and fully censored datasets. Our experiments demonstrate that fairness alignment can be easily undermined through RAG without the need for fine-tuning or retraining. Even with fully censored and supposedly unbiased external datasets, RAG would still lead to biased outputs. Our findings underscore the limitations of current alignment methods in the context of RAG-based LLMs and highlight the urgent need for new strategies to ensure fairness. We propose potential mitigations and call for further research to develop robust fairness safeguards in RAG-based LLMs.
Anthology ID:
2025.findings-emnlp.984
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2025
Month:
November
Year:
2025
Address:
Suzhou, China
Editors:
Christos Christodoulopoulos, Tanmoy Chakraborty, Carolyn Rose, Violet Peng
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
18145–18170
Language:
URL:
https://aclanthology.org/2025.findings-emnlp.984/
DOI:
Bibkey:
Cite (ACL):
Mengxuan Hu, Hongyi Wu, Ronghang Zhu, Zihan Guan, Dongliang Guo, Daiqing Qi, and Sheng Li. 2025. No Free Lunch: Retrieval-Augmented Generation Undermines Fairness in LLMs, Even for Vigilant Users. In Findings of the Association for Computational Linguistics: EMNLP 2025, pages 18145–18170, Suzhou, China. Association for Computational Linguistics.
Cite (Informal):
No Free Lunch: Retrieval-Augmented Generation Undermines Fairness in LLMs, Even for Vigilant Users (Hu et al., Findings 2025)
Copy Citation:
PDF:
https://aclanthology.org/2025.findings-emnlp.984.pdf
Checklist:
 2025.findings-emnlp.984.checklist.pdf