Papers

论文提出 Personal Imitation Game:让机器在对话中模仿特定个人

Who are you? The Personal Imitation Game

Why we picked it

这篇论文把图灵测试升级成了个人版:用十年聊天记录训练出一个模仿特定真人的系统。20 人盲测里,审讯者分不清它和真人朋友。做智能体、数字分身或关注隐私边界的人都该看看。

arXiv 论文《Who are you? The Personal Imitation Game》提出比图灵 1950 年模仿游戏更难的测试:机器能否在对话中复现一个具体的人,而非泛化的人类。研究以 30 名参与者建立人类基线,并梳理审讯者依赖的身份信号,包括文体特征、人格特质和生平事实。系统演示方面,团队基于一位论文作者十年个人数字通信记录构建的自动民族志数据集,设计了一套 LLM 系统,让其为这位作者“代言”。

Body

在 20 人参与的欺骗研究中,审讯者在 10 轮里有 7 轮选择该系统而非作者本人的朋友;作者本人则在 20 次中有 13 次被选中,两个比例在该样本量下均与随机猜测无法区分。论文还讨论了个人 AI 智能体的应用前景,以及这类“个人级模仿”对身份认定和隐私带来的影响。

Original · arXiv cs.AI

Who are you? The Personal Imitation Game

Large Language Models (LLMs) now perform strongly on Alan Turing's 1950 Imitation Game, so we propose a harder challenge, the Personal Imitation Game, which asks whether a machine can replicate a specific individual in conversation rather than a generic human. Testing with 30 participants, we establish a human baseline and report the identity signals interrogators rely on: stylometric features, personality traits, and biographical facts. To demonstrate the framework, we design an LLM-based system that plays the game for one of the paper's authors, drawing on an autoethnographic dataset spanning ten years of personal digital communication. In a 20-person deception study, interrogators choose our system over the author's own friends in 7 of 10 rounds, and the author over our system in 13 of 20; neither rate is distinguishable from chance at this sample size. We consider design improvements, applications of personal AI agents, and implications for identity and privacy.

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include LLM, Turing test, agents, privacy.

Source attribution: curated from arXiv cs.AI (@Muhammad Rizky Wellyanto, Hajara-Yasmin Isa, Kexin Xue, Vishal Ravichandran, Wenpeng Lu, Jerry O. Talton, Ranjitha Kumar).