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

October 10, 2024 Β· Declared Dead Β· πŸ› Findings of ACL. EMNLP. Conference on Empirical Methods in Natural Language Processing

πŸ‘» CAUSE OF DEATH: Ghosted
No code link whatsoever

"No code URL or promise found in abstract"

Evidence collected by the PWNC Scanner

Authors Mengxuan Hu, Hongyi Wu, Zihan Guan, Ronghang Zhu, Dongliang Guo, Daiqing Qi, Sheng Li arXiv ID 2410.07589 Category cs.IR: Information Retrieval Cross-listed cs.CL Citations 12 Venue Findings of ACL. EMNLP. Conference on Empirical Methods in Natural Language Processing Last Checked 4 months ago
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 the external dataset. 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 can 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.
Community shame:
Not yet rated
Community Contributions

Found the code? Know the venue? Think something is wrong? Let us know!

πŸ“œ Similar Papers

In the same crypt β€” Information Retrieval

Died the same way β€” πŸ‘» Ghosted