China’s use of generative artificial intelligence in propaganda and influence operations is drawing increasing scrutiny, with new reports pointing to more sophisticated methods aimed at shaping narratives abroad, creating fake identities, and targeting political figures.
A detailed account published by The Diplomat highlights how Chinese-linked operations are harnessing AI tools to refine content laundering, spread state propaganda, and run smear campaigns through fabricated online personas. Analysts warn that by tailoring content to local languages and cultures, these campaigns could embed Beijing-aligned narratives more deeply, particularly among younger audiences in developing regions where social media is a dominant source of information.
AI-enabled influence networks
Evidence of China’s growing reliance on AI in influence operations surfaced in early August, when two Vanderbilt University professors released findings based on Chinese documents linked to the private firm GoLaxy.
The documents showed AI being deployed not only to generate misleading content targeting audiences in Hong Kong and Taiwan, but also to collect information on US lawmakers. These profiles, experts noted, could be exploited later in espionage or lobbying efforts. Several major technology companies have also tracked the trend. OpenAI, Meta, and the research firm Graphika have all documented Chinese-linked actors using AI to strengthen foreign propaganda networks.
Graphika’s Falsos Amigos report, published last month, identified 11 fake news websites set up between December 2024 and March 2025. These sites used AI-generated images and logos to build credibility and pushed Beijing-friendly narratives across multiple languages.
Tactics exposed
OpenAI’s June threat report described how banned ChatGPT accounts were used to generate fake news outlet names, logos, and social media personas. Among them was a campaign called “Uncle Spam,” in which AI-created profiles posed as US military veterans critical of the Trump administration.
These personas, supported by AI-generated comments, were designed to intensify political divisions in the United States. A key element of these operations has been the simulation of organic engagement. Investigators found coordinated activity in which a “main” account posted controversial remarks, followed by a wave of AI-generated replies mimicking debate among users. Such tactics were used to comment on US tariffs, presenting both support and opposition to appear authentic.
Smear campaigns against activists
Beyond shaping political narratives, AI-driven operations have also been deployed against individual critics of Chinese policies. Meta traced a TikTok account and a Facebook page to a campaign targeting Pakistani activist Mahrang Baloch, who has spoken against Chinese projects in Balochistan. A falsified video accused her of appearing in pornography, while hundreds of AI-generated comments in English and Urdu were posted to amplify the smear.
Growing concerns
These findings underscore the growing sophistication of Beijinglinked disinformation strategies. While earlier campaigns relied on deepfakes or isolated false personas, the latest evidence suggests a coordinated push to build entire ecosystems of fabricated media outlets, supported by AI tools that enhance reach and authenticity. Analysts argue that such operations demand urgent responses from social media platforms, software developers, and governments. Left unchecked, they warn, the integration of AI into state-backed propaganda could entrench disinformation more effectively than previous efforts and erode trust in the digital information space worldwide