The Imitation Game: Using Large Language Models as Chatbots to Combat Conversational Cybercrime
The Imitation Game: Using Large Language Models as Chatbots to Combat Chat-Based Cybercrimes
AI summary
The researchers propose the LURE (LLM-based User Response Engagement) system, which embeds a large language model as an active agent in adversarial chat environments to identify and counter conversational cybercrime. In an illegal video chat scam scenario on Telegram, the system engaged 53 actors across 98 groups, maintaining multi-turn conversations in over 56% of interactions without being identified as a bot. LURE combines automated discovery, adversarial interaction, and OCR-based analysis of image payment data, and reveals scam behavior patterns such as payment flows, upsell strategies, and platform migration.