AI 生成文本具有一致的计量文体足迹，但 AI 编辑文本不具有
AI Writers Have a Consistent Stylometric Footprint, but AI Editors Do Not
- Boston University（波士顿大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
该研究发现 AI 生成文本具有一致的计量文体足迹，而 AI 编辑文本没有，二者留下的痕迹存在差异，需分别研究。
AI中文摘要：
已有研究表明，大型语言模型（LLM）生成的文本在计量文体学上与人类撰写的文本存在差异，相关研究包括 Andre 等人（2023）、Shah 等人（2023）、Opara 等人（2024）、Soto（2024）、Li 等人（2025）以及 Selvioglu 等人（2025）的工作。不过，LLM 如今不仅用于生成文本，还用于编辑人类撰写的文本，目前尚不明确这两种操作是否会留下相同的痕迹。本文研究显示，AI 生成文本会留下一致的“计量文体足迹”：在 8 种 LLM 和 5 个领域中，一小部分特征（主要是熵和词汇多样性）能够始终如一地将 AI 生成文本与人类撰写文本区分开来，其余特征则高度依赖于具体领域和生成器。然而，AI 编辑文本并不会产生相同的足迹。与人类撰写的源文本相比，AI 编辑文本仅表现出词汇多样性的小幅提升和熵的下降，而非 AI 生成文本所特有的二者同时提升的情况。原本对生成作用不大的词汇密度，反而成为与编辑相关的主导信号。因此，计量文体特征能够区分 AI 编辑文本与 AI 生成文本，但在区分 AI 编辑文本与人类撰写文本时效果要差得多。研究结果表明，“AI 文本”并非单一现象：生成与编辑会留下性质不同的计量文体痕迹，应当分别加以研究。
英文摘要：
Text generated by large language models (LLMs) has been shown to be stylometrically distinct from human-written text (Andre et al., 2023; Shah et al., 2023; Opara, 2024; Soto et al., 2024; Li and Zhang, 2025; Selvioglu et al., 2025). But LLMs are increasingly used not only to generate text but also to edit human writing, and it is unclear whether the two leave the same trace. We show that AI generation leaves a consistent "stylometric footprint": a small subset of features, primarily entropy and lexical diversity, consistently separates AI-generated text from human writing across 8 LLMs and 5 domains, while the remaining features depend heavily on the domain and generator. AI editing, however, does not reproduce the same footprint. Relative to their human- written sources, AI-edited texts show only a small increase in lexical diversity and a decrease in entropy, rather than the joint increase that characterizes AI generation. Lexical density, which contributes little to generation, instead becomes the dominant editing-associated signal. Stylometric features therefore separate AI-edited text from AI-generated text but are substantially less effective at separating it from human-written text. Our results suggest that "AI text" is not a single phenomenon: generation and editing leave qualitatively different stylometric traces and should be studied separately.
