Papers for

privacy technology teams

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Large language models try to imitate specific people in chat

Who are you? The Personal Imitation Game

Abstract: 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.

Wed 18 NovHuman-Computer InteractionArtificial IntelligenceComputation and Language
The gist
Instead of asking if a computer can act like any person in conversation, this paper asks if it can imitate one particular person. The authors tested a new challenge where a machine uses personal writing and chats collected over ten years to mimic one of the paper's authors. Human judges often couldn’t tell if they were talking to the real person or the computer system, showing the machine could pass as that individual sometimes. The work also looks at how this technology might affect privacy and the idea of personal identity.
Open → 2611.24888v1