France has identified a network of propaganda sites linked to Beijing.
On June 4, the French foreign digital interference watchdog, Viginum, revealed a network of 13 fake news sites tied to the Chinese state media outlet CGTN.
Dubbed "Fawn Mianju," the operation was initially
时政
(
twitter.com
)
France has identified a network of propaganda sites linked to Beijing.
On June 4, the French foreign digital interference watchdog, Viginum, revealed a network of 13 fake news sites tied to the Chinese state media outlet CGTN.
Dubbed "Fawn Mianju," the operation was initially exposed in 2025. It utilized fake websites and social media accounts to relentlessly push CGTN content to young audiences in multiple languages, including English, French, and Spanish.
The pseudo-articles aggressively promoted China as a global leader in technology and ecology, advocated for French alignment with Beijing, and criticized Western reporting on the treatment of Uyghurs.
After a months-long investigation, Viginum experts successfully traced these 13 sites directly back to CGTN.
The network relied on domain names registered in Beijing and expensive distributed server architectures, indicating an actor with significant financial resources.
Investigators uncovered a test page mimicking the CGTN website, where the administrator accidentally left traces of his login credentials. He was identified as a senior IT project manager at CGTN Digital.
His GitHub profile revealed work on text-generation algorithms and large language models (LLMs). Since late 2025, over 2,300 CGTN articles were quickly rewritten by AI and republished on the fake network to target a younger demographic.
Stylometric analysis confirmed the suspected use of AI, showing a drastic reduction in sentence variation compared to the human-written text on the original CGTN site.
The automated nature of the operation was further evidenced by the rapid turnaround time, with articles consistently republished less than an hour after first appearing on CGTN.
Despite buying visibility to target 89 countries—primarily in French-speaking Africa—the network struggled for traction, never exceeding 15,000 views and relying heavily on artificial engagement from accounts in Burundi.
While ultimately deemed an operational failure due to its low reach, the campaign exposes China’s active use of AI automation to clandestinely push state narratives to Western and African youth.
https://t.co/W5cA1a6W1f
点击图片查看原图
