Personalized Attacks of Social Engineering in Multi-turn Conversations -- LLM Agents for Simulation and Detection

2025-03-18

AI Summary

To address the challenge of detecting personalized social engineering (SE) attacks in multi-turn social media dialogues, this paper proposes a personality-driven dynamic defense framework. First, we introduce SE-VSim—a large language model (LLM)-based agent simulator that generates realistic multi-turn SE attacks and models victim psychological susceptibility using the Big Five personality framework. Second, we design SE-OmniGuard, an interpretable, real-time detection and response system integrating personality profiling, behavioral sequence monitoring, and attack strategy assessment. This work pioneers a tightly coupled simulation-and-defense paradigm that jointly leverages personality modeling and multi-turn dialogue monitoring. Evaluated on over 1,000 simulated dialogues across diverse attacker roles—including recruiter, agency, and journalist—SE-OmniGuard achieves a 27.4% improvement in F1-score over baseline methods, demonstrating robust cross-scenario adaptability and operational efficacy.

Technology Category

• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior
• Multiagent Systems: Adversarial Agents
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

The rapid advancement of conversational agents, particularly chatbots powered by Large Language Models (LLMs), poses a significant risk of social engineering (SE) attacks on social media platforms. SE detection in multi-turn, chat-based interactions is considerably more complex than single-instance detection due to the dynamic nature of these conversations. A critical factor in mitigating this threat is understanding the mechanisms through which SE attacks operate, specifically how attackers exploit vulnerabilities and how victims' personality traits contribute to their susceptibility. In this work, we propose an LLM-agentic framework, SE-VSim, to simulate SE attack mechanisms by generating multi-turn conversations. We model victim agents with varying personality traits to assess how psychological profiles influence susceptibility to manipulation. Using a dataset of over 1000 simulated conversations, we examine attack scenarios in which adversaries, posing as recruiters, funding agencies, and journalists, attempt to extract sensitive information. Based on this analysis, we present a proof of concept, SE-OmniGuard, to offer personalized protection to users by leveraging prior knowledge of the victims personality, evaluating attack strategies, and monitoring information exchanges in conversations to identify potential SE attempts.

Problem

• Simulate social engineering attacks using LLM agents
• Assess personality traits' impact on susceptibility to manipulation
• Develop personalized protection against multi-turn SE attacks

Innovation

• LLM-agentic framework for SE attack simulation
• Modeling victim agents with varying personality traits
• Personalized protection leveraging victim personality knowledge