In a sophisticated revelation that highlights the evolving nature of digital statecraft, the French government agency VIGINUM recently pulled back the curtain on a vast, AI-driven disinformation network. Spearheaded by this agency—tasked specifically with shielding France from foreign digital interference—the operation, codenamed “Fawn Mianju,” successfully unmasked a constellation of counterfeit news websites. These outlets, which emerged in mid-2025, were carefully crafted to mimic legitimate journalism, but their true purpose was to serve as a mouthpiece for Beijing. By posing as independent outlets in French, English, Spanish, and Vietnamese, these sites gained a veneer of credibility while functioning essentially as a digital extension of the China Global Television Network (CGTN), the state-controlled propaganda arm of the Chinese Communist Party.
The narrative strategy behind these sites was as calculated as it was insidious, specifically targeting the digital habits of young people in Western nations and French-speaking regions of Africa. Rather than adopting typical, dry political rhetoric, the network focused on topics that appeal to Gen Z and younger demographics. The content praised China’s strides in aerospace and artificial intelligence, positioned the nation as a visionary leader for the “Global South,” and emphasized its commitment to environmental causes. By wrapping these messages in a modern, youth-oriented aesthetic, the network aimed to normalize pro-China sentiment, even going so far as to actively counter Western reporting on human rights abuses, such as the treatment of Uyghurs, by framing independent investigations as misguided or hostile propaganda.
Technologically, the operation was a marvel of modern efficiency, utilizing large language models (LLMs) to weaponize content at an unprecedented scale. Investigations revealed that the network did not merely report news; it functioned as a parasitic extension of CGTN. Analysts found that out of roughly 3,000 articles published by CGTN in late 2025, over 2,300 were scraped, lightly rephrased, and reposted to the fake network within less than an hour of the original publication. To ensure this flood of content remained consistent in tone, the operators relied on AI tools to rewrite the copy, adapting the syntax to suit the preferences of a younger audience. Stylometric analysis conducted by French investigators confirmed the presence of AI; the distinct lack of natural variation in sentence length and punctuation revealed a statistical signature that no human writer would produce on their own.
Linking this digital labyrinth to its source required sophisticated investigative detective work. Beyond the content itself, French authorities followed the proverbial paper trail. The domain names for these fake outlets were registered in Beijing and hosted on Alibaba Cloud, suggesting a project with significant state resources behind it—far beyond what an independent blogger could afford. The “smoking gun” was discovered when an administrator for one of the primary French-language sites, “Actu Méridien,” inadvertently left behind digital breadcrumbs. By tracing these login credentials, investigators identified a senior project manager at CGTN Digital. A subsequent review of the individual’s GitHub profile uncovered extensive work on integrating LLMs into web infrastructure, effectively linking the high-level propaganda strategy directly to the technical implementation of the disinformation chain.
To maximize their influence, the network extended its reach beyond static web pages, engaging in coordinated campaigns across social media platforms like Facebook and Threads. These campaigns, while wide-reaching—spanning users across 89 countries—saw varying levels of success. Despite the sophisticated automation under the hood, much of the engagement appeared inauthentic, with pockets of significant, unnatural activity concentrated in countries like Burundi. While some reports garnered thousands of views, the overall footprint of the project remained somewhat stunted, leading to the gradual inactivity of several key sites, including Actu Méridien, as the campaign lost momentum or faced increased scrutiny from global cybersecurity firms like Graphika, which had begun identifying these pages earlier in the year.
Ultimately, the VIGINUM investigation serves as a sobering reminder of how AI is transforming the geopolitical propaganda landscape. By leveraging sophisticated tools to manufacture consent and reshape the narratives consumed by the next generation, states are finding new, covert ways to exert influence across borders. Beijing’s “Fawn Mianju” operation was not just a collection of fake news; it was a blueprint for the future of information warfare. As nations look to protect their digital spheres, the case highlights the urgent need for critical media literacy and the importance of agencies like VIGINUM in verifying the authenticity of information in an era where the line between news and algorithmically generated propaganda has become dangerously thin.