IMPersona: Evaluating Individual Level LM Impersonation

Quan Shi, Carlos E Jimenez, Stephen Dong, Brian Seo, Caden Yao, Adam Kelch, Karthik R Narasimhan

Keywords: Impersonation, Language Models, Personalization, Stylistic Mimicry, Contextual Knowledge, AI Evaluation, Social Engineering, Ethical AI, Memory-Augmented Models, Human-AI Interaction

TL;DR: We train LLMs to impersonate individuals by mimicking style and personal knowledge, surpassing prompting methods, while raising safety and alignment concerns.

Abstract: As language models achieve increasingly human-like capabilities in conversational text generation, a critical question emerges: to what extent can these systems simulate the characteristics of specific individuals? To evaluate this, we introduce IMPersona, a framework for evaluating LMs at impersonating specific individuals' writing style and personal knowledge. Using supervised fine-tuning and a hierarchical memory-inspired retrieval system, we demonstrate that even modestly sized open-source models, such as Llama-3.1-8B-Instruct, can achieve impersonation abilities at concerning levels. In blind conversation experiments, participants (mis)identified our fine-tuned models, with memory augmentation, in 44.44\% of interactions, compared to just 25.00\% for the best prompting-based approach. We analyze these results to propose detection methods and defense strategies against such impersonation attempts. Our findings raise important questions about both the potential applications and risks of personalized language models, particularly regarding privacy, security, and the ethical deployment of such technologies in real-world contexts.

Track: ML track

Submission Number: 6