分析AI生成宣传内容的制作链条，发现其使用特定指令模板并混合多个大模型。
Propaganda Forensics: Recovering the Generation Pipeline of an AI-Driven Influence Campaign
- 通过对比真实新闻与宣传文，识别出模糊、情绪化和少引用等典型洗脑特征。
- 在50个网站发现提示词泄露，其中包含完整十点写作规范，导致文章高度重复。
- 检测出内容源自Llama 3和Mistral系列模型，支持多模型协同生成推断。
我们对近期一次由AI驱动的影响行动背后的生成流程进行了法医分析。提出PROPAGIA数据集，包含2646篇来自Storm-1516/CopyCop行动的法语宣传文章，由VIGINUM与INSIKT GROUP于2025年披露；作为对照，采用同一时期的人类撰写主流法语媒体数据集SIPA。通过主题建模、模糊性与情感分析，发现PROPAGIA在模糊性、主观性和负面情绪上显著高于SIPA，且引用来源更少。进一步在84个PROPAGIA网站中的50个中发现提示词指令泄露，包括一份逐字记录的十点编辑规范，解释了部分差异特征，并揭示了跨文章的高度冗余现象。最后，基于重写检测方法，验证了INSIKT GROUP对Llama 3系列模型的归因，同时暗示了Mistral系列模型的参与。
原文摘要 · Abstract (English)
We present a forensic analysis of the generation pipeline behind a recent AI-driven influence campaign. We introduce PROPAGIA, a corpus of 2,646 propagandist French articles from the Storm-1516/CopyCop campaign disclosed by VIGINUM and INSIKT GROUP in 2025. For comparison, we rely on SIPA, a corpus of human-written French mainstream press from the same period. Using topic modeling, vagueness and sentiment analysis, we first isolate persuasion techniques characteristic of propaganda, with PROPAGIA far exceeding SIPA in vagueness, subjectivity and negativity, and citing fewer sources. We then find prompt instruction leaks on 50 of the 84 PROPAGIA websites, including a verbatim ten-point editorial specification accounting for several of these differences, together with high cross-article redundancy. Finally, we show that rewriting-based detection supports INSIKT GROUP's attribution to the Llama 3 family, but also suggests the involvement of Mistral-family models.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。