Can LLMs be Scammed? A Baseline Measurement Study

TL;DR

This study systematically evaluates the vulnerability of large language models to various scam tactics using a benchmark based on the FINRA taxonomy, revealing susceptibility patterns and informing future improvements.

Contribution

The paper introduces a structured benchmark for assessing LLMs' scam detection capabilities across diverse scenarios and analyzes model vulnerabilities in detail.

Findings

GPT-4 outperforms GPT-3.5 and Llama in scam detection

Certain scam tactics are more effective against LLMs

Model susceptibility varies with persona traits and persuasive techniques

Abstract

Despite the importance of developing generative AI models that can effectively resist scams, current literature lacks a structured framework for evaluating their vulnerability to such threats. In this work, we address this gap by constructing a benchmark based on the FINRA taxonomy and systematically assessing Large Language Models' (LLMs') vulnerability to a variety of scam tactics. First, we incorporate 37 well-defined base scam scenarios reflecting the diverse scam categories identified by FINRA taxonomy, providing a focused evaluation of LLMs' scam detection capabilities. Second, we utilize representative proprietary (GPT-3.5, GPT-4) and open-source (Llama) models to analyze their performance in scam detection. Third, our research provides critical insights into which scam tactics are most effective against LLMs and how varying persona traits and persuasive techniques influence these vulnerabilities. We reveal distinct susceptibility patterns across different models and scenarios, underscoring the need for targeted enhancements in LLM design and deployment.