GPT-4 is judged more human than humans in displaced and inverted Turing tests
This highlights a critical challenge for everyday AI detection in informal settings, showing that current methods are unreliable when not actively interrogating, which is incremental but urgent for online conversation safety.
The study investigated how well humans and large language models (LLMs) can distinguish between humans and AI in online conversations using modified Turing tests, finding that both AI and displaced human judges performed below chance accuracy and judged GPT-4 as human more often than actual humans.
Everyday AI detection requires differentiating between people and AI in informal, online conversations. In many cases, people will not interact directly with AI systems but instead read conversations between AI systems and other people. We measured how well people and large language models can discriminate using two modified versions of the Turing test: inverted and displaced. GPT-3.5, GPT-4, and displaced human adjudicators judged whether an agent was human or AI on the basis of a Turing test transcript. We found that both AI and displaced human judges were less accurate than interactive interrogators, with below chance accuracy overall. Moreover, all three judged the best-performing GPT-4 witness to be human more often than human witnesses. This suggests that both humans and current LLMs struggle to distinguish between the two when they are not actively interrogating the person, underscoring an urgent need for more accurate tools to detect AI in conversations.