The Chinese government’s external-propaganda efforts are receiving a boost from AI. A report published by Graphika last month analyzed what it described as a China-based network of domains and social media accounts using AI tools to “launder” Chinese state-media content and disseminate it in different languages. Here are the report’s key findings:

Graphika identified a network of 11 domains and 16 companion social media accounts that laundered exclusively English-language articles originally published by the Chinese state media outlet CGTN. The identified accounts are active on Facebook, Instagram, Mastodon, Threads, and X.

The assets almost certainly used AI tools to translate and summarize articles from CGTN, likely in an attempt to disguise the content’s origin.

The network assets disseminated primarily pro-China, anti-West content in English, French, Spanish, and Vietnamese. They also utilized AI tools to generate logos and text specifically designed to target different audiences, including young audiences in Africa, the Americas, Asia, and Europe.

The network’s social media accounts failed to receive any organic traction. On Facebook, we identified pages that are part of the network and ran ads to promote their content and increase their visibility. On X, posts from the associated accounts appeared in top search results for some topics.

We could not attribute these domains and their related accounts to a specific actor(s) based on open-source information. We assess that the actor(s) behind the activity are very likely located in China based on the technical indicators of the domains and social media accounts. [Source]

(In our Borrowed Boats report from last year, CDT tracked similar phenomena of “embedded propaganda,” or Chinese state-media content reproduced in African media outlets, which at times appeared without proper attribution to the original source.)

These sort of campaigns of inauthentic activity are regularly removed from Western social media platforms. According to its most recent update, in the second quarter of this year, Google’s Threat Analysis Group removed over 7,700 YouTube channels and other accounts that were part of coordinated influence operations linked to the PRC. The accounts shared content in English and Chinese that discussed U.S. foreign affairs, promoted Xi Jinping and China, and criticized the Philippines. This follows similar account removals implemented over the past few years by tech companies such as Microsoft, Meta, and Facebook in order to counter Chinese online influence operations.

Now, with the help of AI, these influence operations are becoming more complex, effective, and difficult to combat. In February, OpenAI reported that a Chinese-origin network used its ChatGPT AI chatbot to build a tool for compiling and analyzing overseas social media content about sensitive issues and marketing it to Chinese authorities. Last month, The New York Times and researchers from Vanderbilt University reported that the Chinese government is working with Chinese AI companies to monitor and manipulate public opinion both at home and abroad. Internal documents from one such company showed that it had collected data on 117 members of the U.S. Congress and over 2,000 American political figures and thought leaders, and undertaken influence campaigns in Hong Kong and Taiwan. The company, GoLaxy, was founded in 2010 by a research institute at the state-controlled Chinese Academy of Sciences and has worked with top-level Chinese intelligence and military bodies. NYT’s Julian E. Barnes provided more detail on how GoLaxy uses AI for “information warfare”:

A new technology can track public debates of interest to the Chinese government, offering the ability to monitor individuals and their arguments as well as broader public sentiment. The technology also has the promise of mass-producing propaganda that can counter shifts in public opinion at home and overseas.

[…] The new technology allows the Chinese company GoLaxy to go beyond the election influence campaigns undertaken by Russia in recent years, according to the documents.

[…] After being contacted by The Times, GoLaxy began altering its website, removing references to its national security work on behalf of the Chinese government.

[…] Publicly, GoLaxy advertises itself as a firm that gathers data and analyzes public sentiment for Chinese companies and the government. But in the documents, which were reviewed by The Times, the company privately claims that it can use a new technology to reshape and influence public opinion on behalf of the Chinese government.

[…] GoLaxy’s public-facing platform, according to its website, has begun using DeepSeek, an advanced artificial intelligence model developed by a Chinese company. GoLaxy can quickly craft responses that reinforce the Chinese government’s views and counter opposing arguments. Once put into use, such posts could drown out organic debate with propaganda. [Source]

General Paul Nakasone, the former head of the U.S. National Security Agency, recently said that while China has lagged behind Russia in influence operations against the U.S., AI tools may radically change its ability, adding “I do see this as being really the next generation of what we’re going to see in gray zone conflict in the future.” Indeed, the U.S. government is itself reportedly planning to purchase AI tools that create and distribute propaganda overseas in a bid to “influence foreign target audiences,” “suppress dissenting arguments,” and “increase the scale of influence operations.”

The competition over online narratives has spread to many different regions around the world. Earlier this year, the Central American think tank Expediente Abierto and the Latin American digital research media outlet ProBox published a report in Spanish on Chinese digital propaganda in Central America. The report identified influential pro-China narratives in Panama, El Salvador, and Costa Rica, and mapped key actors involved in their dissemination. Global Voices translated an introduction to the report in July, which included the following findings:

China’s soft power in the [Central American] region is no longer limited to cultural or educational expressions. It now operates as a key geopolitical tool in the battle over narratives. Through targeted campaigns, partnerships with local actors, and digital diplomacy, China seeks to shape the Central American information ecosystem to reinforce its legitimacy and counter the influence of rival powers, especially that of the United States. This phenomenon poses concrete challenges for the democratic health of the countries in the region, where freedom of expression and information pluralism are under increasing pressure.

[…] The study shows that China does not apply a one-size-fits-all strategy but rather, adapts its narratives to the political, economic, and cultural context of each country. In all cases, pro-China narratives are primarily promoted by Chinese embassies and diplomats active on social media, Chinese state media like Xinhua and CGTN en Español, local media outlets that uncritically replicate this content, and academic, government, or pro-establishment actors in each country.

The Chinese digital strategy not only seeks to reinforce a positive presence but also to minimize criticism of its authoritarian model, obscure human rights violations, and present its political system as a successful alternative to Western liberalism. This is framed as a form of “unconditional cooperation,” which resonates strongly in authoritarian or semi-authoritarian contexts. [Source]

AI tools complement other forms of influence operations carried out by the Chinese government. Reports have shown that Chinese consulates in New York and other American cities engage in gray-zone intelligence operations to influence local politics. This week, The New York Times published an investigation into how the Chinese Consulate in New York influences the city’s elections.

On the domestic front, AI has been instrumentalized for various propaganda purposes and other ends to enhance government control. Zhang Wenjun, a deputy director at Beijing’s Party School, recently argued that generative AI can play a key role in getting individuals to follow Party ideology, as Lingua Sinica’s latest China Chatbot column noted. And Chinese authorities have already used AI to help censor commemorations of the Tiananmen Massacre, monitor public opinion online, and surveil users of Telegram and VPN services.