Detecting Sentiment Steering Attacks on RAG-enabled Large Language Models

2026-03-17

AI Summary

This study addresses the vulnerability of Retrieval-Augmented Generation (RAG) systems to sentiment manipulation attacks, wherein adversaries induce large language models to produce responses with targeted emotional biases by tampering with retrieved content. Focusing specifically on the underexplored dimension of affective safety in RAG architectures, this work proposes a novel detection framework grounded in the emotional consistency between retrieved passages and generated outputs. By integrating sentiment analysis, anomaly detection, and attention mechanisms, the framework effectively identifies subtle sentiment deviations indicative of manipulation. Experimental results demonstrate that the approach achieves high precision in detecting diverse sentiment manipulation attacks on standard benchmarks, substantially enhancing the robustness and trustworthiness of RAG systems while filling a critical gap in the affective security of large language models.

Technology Category

• Natural Language Processing: Safety and Robustness
• Cognitive Modeling & Cognitive Systems: Affective Computing
• Computer Vision: Adversarial Attacks & Robustness

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

The proliferation of large-scale IoT networks has been both a blessing and a curse. Not only has it revolutionized the way organizations operate by increasing the efficiency of automated procedures, but it has also simplified our daily lives. However, while IoT networks have improved convenience and connectivity, they have also increased security risk due to unauthorized devices gaining access to these networks and exploiting existing weaknesses with specific attack types. The research proposes two lightweight deep learning (DL)-based intelligent intrusion detection systems (IDS). to enhance the security of IoT networks: the proposed convolutional neural network (CNN)-based IDS and the proposed long short-term memory (LSTM)-based IDS. The research evaluated the performance of both intelligent IDSs based on DL using the CICIoT2023 dataset. DL-based intelligent IDSs successfully identify and classify various cyber threats using binary, grouped, and multi-class classification. The proposed CNN-based IDS achieves an accuracy of 99.34%, 99.02% and 98.6%, while the proposed LSTM-based IDS achieves an accuracy of 99.42%, 99.13%, and 98.68% for binary, grouped, and multi-class classification, respectively.

Problem

• IoT security
• intrusion detection
• cyber threats
• unauthorized access
• network attacks

Innovation

• lightweight deep learning
• intrusion detection system
• CNN-based IDS
• LSTM-based IDS
• IoT security