Researchers Share EMNLP2026 Findings on AI Text Stylometrics
Announcement highlights consistent patterns in machine-generated content versus edited versions.
TLDR
Zhengyang Shan, a PhD student at Boston University, posted about an EMNLP2026 main paper he co-led with Yukyung Lee. The announcement states that AI-generated text maintains a consistent stylometric footprint across models and domains, while AI-edited text shows a different pattern. Najoung Kim, an assistant professor of linguistics at Boston University, reposted the announcement. The post links to the paper on arXiv and includes a research slide image. No additional details on methods or results appear in the visible posts.
Combined views
2.9K
3 Sources, first seen 26d ago