A Chinese AI persona army has been exposed through leaked internal documents from a Beijing-based company, GoLaxy, revealing sophisticated information warfare tactics targeting American public figures and citizens. On 19 September 2025, The Record reported that researchers discovered a cache of 399 pages detailing how China uses artificial intelligence to create realistic digital personas designed to infiltrate social media and manipulate public opinion. The article begins:
On a jet-lagged April morning earlier this year, Brett Goldstein’s phone buzzed with a message from a trusted colleague asking him to check out a link he’d just found. Goldstein, a former government technologist and now a special adviser at Vanderbilt University’s Institute for National Security, was in the middle of a conference. Later, back in his hotel room, he scrolled through to the message again. The note was insistent and included a link. “I hate links,” he said. Still, he was curious, so he scanned the link for malware, ran a few checks, and clicked — and what appeared on his screen was a trove of PDFs all written in Mandarin. There were pages of dense Mandarin characters, something that looked like technical schematics, and then, strangely, photos of prominent Americans. “At first I thought it was spam,” Goldstein remembered. But as he scrolled through the files, he realized it was something else: a leak.
Read more: https://therecord.media/golaxy-china-artificial-intelligence-papers
Key Points
- GoLaxy documents revealed dossiers on 2,000 American public figures and 117 Republican Congress members
- Chinese company employs nearly 1,000 people and maintains ties to military and civilian intelligence
- AI personas use China’s DeepSeek technology to create highly realistic digital identities for influence operations
- GoLaxy began deleting website content after New York Times contacted them for comment
China’s Artificial Intelligence Influence Operations: How Beijing Weaponizes AI to Manipulate Global Information
China has systematically weaponized artificial intelligence to revolutionize its influence operations, deploying sophisticated AI-powered strategies that merge psychological profiling, algorithmic targeting, and mass surveillance to manipulate global information environments with unprecedented precision. China’s military developed “algorithmic cognitive warfare” using AI to create detailed psychological profiles for targeted influence operations, with PLA scholars envisioning personalized content generation at scale that exploits recommendation algorithms for maximum impact. Chinese threat actor Spamouflage Dragon uses generative AI to create convincing online personas that influence public opinion across global platforms, while China has developed free advanced AI models like DeepSeek and Qwen that rival Western equivalents despite export restrictions, mirroring debt-trap diplomacy by creating technological dependency among nations unable to afford alternatives.
The operational sophistication of Chinese AI influence campaigns has reached alarming levels, with Beijing deploying these capabilities for comprehensive surveillance and disinformation operations targeting democratic societies worldwide. OpenAI banned multiple accounts for using ChatGPT to develop the “Qianyue Overseas Public Opinion AI Assistant,” designed to monitor anti-China protests globally and collect real-time data from X, Facebook, YouTube, Instagram, Telegram, and Reddit for sharing with Chinese authorities.
Stanford research confirms that Chinese institutions are rapidly closing the AI quality gap, with Chinese AI models reaching near parity on key benchmarks after trailing US models by double-digit percentages just one year earlier, while Chinese-origin clusters used AI-generated profile photos to target Myanmar, Taiwan, and Japan with content supporting military juntas and criticizing government ties with the United States.
The scope and scale of Chinese AI-enabled influence operations demonstrate a coordinated state strategy to undermine democratic discourse and expand authoritarian influence through technological superiority. Chinese operatives have leveraged AI tools not only for content creation but also for operational efficiency, with China’s 22 million-strong internet troll army now enhanced by artificial intelligence capabilities that enable automated content generation and coordinated amplification across platforms.
The Diplomat analysis reveals that Chinese AI-driven operations have evolved beyond simple propaganda to sophisticated multi-layered campaigns, including fake French news sites targeting francophone youth and leveraging generative AI to tailor content to local languages and cultural contexts. OpenAI investigators confirm this represents “a growing range of covert operations using a growing range of tactics,” with Chinese actors combining traditional influence techniques with cutting-edge AI capabilities to create false impressions of organic engagement and manipulate public discourse at unprecedented scale.
External References:
- For Beijing’s Foreign Disinformation, the Era of AI-Driven Operations Has Arrived — The Diplomat
- OpenAI takes down covert operations tied to China and other countries — NPR
- U.S.‘s AI lead over China rapidly shrinking, Stanford report finds — Axios
Disclaimer
The Global Influence Operations Report (GIOR) employs AI throughout the posting process, including generating summaries of news items, the introduction, key points, and often the “context” section. We recommend verifying all information before use. Additionally, images are AI-generated and intended solely for illustrative purposes. While they represent the events or individuals discussed, they should not be interpreted as real-world photography.
PREVIOUS POST