Are You Human? An Adversarial Benchmark to Expose LLMs
This addresses the critical need for reliable, real-time LLM detection to prevent misuse in scams and deception, though it is incremental as it builds on existing detection methods with a new benchmark.
The paper tackles the problem of detecting LLMs impersonating humans in conversation by developing a benchmark with explicit and implicit challenges, achieving 78.4% detection success with explicit challenges and 22.9% with implicit ones, and validating real-world applicability through user studies.
Large Language Models (LLMs) have demonstrated an alarming ability to impersonate humans in conversation, raising concerns about their potential misuse in scams and deception. Humans have a right to know if they are conversing to an LLM. We evaluate text-based prompts designed as challenges to expose LLM imposters in real-time. To this end we compile and release an open-source benchmark dataset that includes 'implicit challenges' that exploit an LLM's instruction-following mechanism to cause role deviation, and 'exlicit challenges' that test an LLM's ability to perform simple tasks typically easy for humans but difficult for LLMs. Our evaluation of 9 leading models from the LMSYS leaderboard revealed that explicit challenges successfully detected LLMs in 78.4% of cases, while implicit challenges were effective in 22.9% of instances. User studies validate the real-world applicability of our methods, with humans outperforming LLMs on explicit challenges (78% vs 22% success rate). Our framework unexpectedly revealed that many study participants were using LLMs to complete tasks, demonstrating its effectiveness in detecting both AI impostors and human misuse of AI tools. This work addresses the critical need for reliable, real-time LLM detection methods in high-stakes conversations.