{"slug": "how-well-can-large-language-models-impersonate-individual-authors-a-comparative", "title": "How Well Can Large Language Models Impersonate Individual Authors? A Comparative Evaluation", "summary": "A comparative evaluation of six large language models found authorship impersonation remained challenging for most, with Gemini 3 Pro showing the strongest apparent mimicry but only because it extensively copied and reused sentences and phrases from training texts at levels far exceeding human authors, according to the arXiv paper 2611.05771v1. The study tested Gemini 3 Pro, Claude Opus, GPT 5.2, GPT 5.2 Pro, DeepSeek 3.2 and Grok 4.1 Fast on short product review texts using simple and stylometric-guidance prompts, measured by cosine distance to training and embedding-distribution visualizations. Follow-up factor analyses found training set consistency was the most robust predictor of mimicry success while author distinctiveness did not generalize, and the authors say the results matter for forensic analysis of textual evidence involving potential LLM-enabled impersonation.", "body_md": "arXiv:2611.05771v1 Announce Type: new \nAbstract: This study investigates the authorship impersonation capacity of six widely used large language models (LLMs): Gemini 3 Pro, Claude Opus, GPT 5.2, GPT 5.2 Pro, DeepSeek 3.2, and Grok 4.1 Fast. The evaluation was conducted using short product review texts selected to ensure high within author consistency in individual writing style. Two types of in context prompt were designed: a simple prompt with minimal instruction and a complex prompt incorporating explicit stylometric guidance for mimicry. Authorship mimicry was evaluated using a cosine distance based distance to training metric, supplemented by visual analyses of embedding distributions. The results indicate that authorship impersonation remained challenging for most LLMs, with Gemini 3 Pro showing the strongest apparent mimicry performance. However, subsequent overlap analysis revealed that Gemini 3 Pro extensively copied and reused sentences and phrases from the training texts, at levels far exceeding those observed among human authors. Follow up factor analyses showed that training set consistency was the most robust predictor of mimicry success across models, whereas author distinctiveness did not generalise as a reliable predictor. Global embedding visualisations further revealed a consistent macro level pattern: for most models, generated texts occupied regions of the embedding space that were largely separated from the corresponding human training clusters. This indicates limited convergence towards target author style and suggests the presence of model specific lexical and phraseological preferences. The findings highlight the importance of distinguishing genuine stylistic imitation from direct textual reuse and have implications for the forensic analysis of textual evidence involving potential LLM enabled impersonation.", "url": "https://wpnews.pro/news/how-well-can-large-language-models-impersonate-individual-authors-a-comparative", "canonical_source": "https://arxiv.org/abs/2611.05771", "published_at": "2026-11-05 05:00:00+00:00", "updated_at": "2026-11-05 05:19:38.364726+00:00", "lang": "en", "topics": ["large-language-models", "ai-research", "ai-safety", "ai-ethics"], "entities": ["Gemini 3 Pro", "Claude Opus", "GPT 5.2", "GPT 5.2 Pro", "DeepSeek 3.2", "Grok 4.1 Fast", "arXiv"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/how-well-can-large-language-models-impersonate-individual-authors-a-comparative", "markdown": "https://wpnews.pro/news/how-well-can-large-language-models-impersonate-individual-authors-a-comparative.md", "text": "https://wpnews.pro/news/how-well-can-large-language-models-impersonate-individual-authors-a-comparative.txt", "jsonld": "https://wpnews.pro/news/how-well-can-large-language-models-impersonate-individual-authors-a-comparative.jsonld"}}