# People cannot distinguish GPT-4 from a human in a Turing test Source: https://arxiv.org/abs/2405.08007 Kiwi News: https://news.kiwistand.com/stories/People-cannot-distinguish-GPT-4-from-a-human-in-a-Turing-test?index=0x66472eb621274ff472652e9c99d89ed36db2a3da47f69030b5dae099888e172b9e3a7968 Submitted by frotograf.eth on 2024-05-17 UTC, 2 upvotes, 0 comments ## Summary In a preregistered randomized Turing test, Cameron R. Jones and Benjamin K. Bergen report that GPT-4 passes as human, a first robust empirical demonstration that an artificial system passes an interactive two-player Turing test. In 5-minute conversations, GPT-4 was judged to be human 54% of the time, outperforming ELIZA at 22% but falling short of actual humans at 67%. The authors suggest that AI deception may go undetected and that stylistic and socio-emotional factors matter more than traditional notions of intelligence in passing the test. ## Article People cannot distinguish GPT-4 from a human in a Turing test Authors: Cameron R. Jones, Benjamin K. Bergen We evaluated 3 systems (ELIZA, GPT-3.5 and GPT-4) in a randomized, controlled, and preregistered Turing test. Human participants had a 5 minute conversation with either a human or an AI, and judged whether or not they thought their interlocutor was human. GPT-4 was judged to be a human 54% of the time, outperforming ELIZA (22%) but lagging behind actual humans (67%). The results provide the first robust empirical demonstration that any artificial system passes an interactive 2-player Turing test. The results have implications for debates around machine intelligence and, more urgently, suggest that deception by current AI systems may go undetected. Analysis of participants' strategies and reasoning suggests that stylistic and socio-emotional factors play a larger role in passing the Turing test than traditional notions of intelligence.