Recently, the Minister of State Security of the Chinese Communist Party, Chen Yixin, has published an article addressing the potential political, network, data, technology, social, and military risks that artificial intelligence may bring. He emphasized the importance of incorporating these risks into the national security system, with “political security” being the top priority. Scholars believe that this ranking reveals that the primary concern for the CCP is not the harm that artificial intelligence may cause to the general public, but rather the potential for information dissemination to bypass the party’s control, challenging its propaganda, censorship, and stability maintenance system.
Chen Yixin’s article titled “Comprehensively Building a Solid Barrier for AI Security to Promote the Healthy and Orderly Development of AI” was published on September 13 by the “China Internet Information” magazine’s WeChat public account. The article noted that artificial intelligence has become a major area of global technological competition and a new battleground for major powers, emphasizing the need for China to establish a comprehensive system for preventing and controlling AI security risks, with CCP leadership guiding the governance of AI throughout the process.
The first risk highlighted by Chen Yixin is the impact of artificial intelligence on the “political security environment.” He warns that “hostile forces” at home and abroad, as well as individuals with ulterior motives, might exploit deepfakes, AI-generated images and text, and so-called “smart internet armies” to create political rumors, spread harmful information, and incite divisive sentiments, launching public opinion warfare and cognitive warfare against China.
Scholar Wu Xizhao (pseudonym) from Sun Yat-sen University in Guangdong told Dajiyuan that the CCP’s definition of AI security differs from that of the United States. While the U.S. focuses on threats to human life posed by AI, the CCP prioritizes regime security. “Their emphasis is on the party’s leadership and the security of its rule. The CCP is not concerned about whether AI will lie, but rather if the information it generates and disseminates can circumvent CCP censorship. They fear that AI may challenge their political narratives.”
Wu Xizhao added that deepfake technology poses risks to public safety, but the CCP defines what qualifies as “rumors,” “harmful information,” and “cognitive warfare.” Criticizing officials, disclosing backgrounds of public figures, or sharing videos of sudden incidents could all be deemed as crossing the political security line.
In recent years, the Minister of State Security in China has rarely publicly published signed articles on specific technology issues. Chen Yixin’s discussion of AI governance as the Minister of State Security has attracted attention from scholars and media both domestically and internationally.
The article also points out that Chinese users processing sensitive information using overseas AI products may lead to massive data leakage domestically. It proposes the establishment of an AI security regulation platform to strengthen risk monitoring, assessment, security review, enforcement mechanisms, and promote the development of specialized laws and regulations for AI research, applications, and governance.
A retired teacher, Mr. Jin, from a university in Jiangsu province, stated that in the past two years, AI products developed under CCP’s guidance are no longer perceived as mere tools but more like propaganda tools. “AI security has become a new tool for the authorities to suppress the thoughts of the common people. Someone recently told me that once AI falls into the hands of the Communist Party, it ceases to be AI and becomes a tool for artificial ignorance. I advise my friends not to use domestically-made AI but to opt for American ones for safety, without the risk of brainwashing.”
Chen Yixin’s article calls for China to accelerate the development of computational power chips, frameworks, and application software, establish national and local data resource pools, and ensure the key technology is “independently controllable.”
A network engineer, Song Yangli (pseudonym), involved in AI research, estimated that tens of thousands of engineers and technicians in China are engaged in AI research. With comprehensive systemic support nationwide, this workforce leverages distilled technology to replicate foreign AI models partially, presenting them as their own research achievements. He stated, “Recently, we’ve been promoting these models to Southeast Asian countries for Malaysia and Indonesia to use Chinese AI models. India will definitely reject them, but it’s uncertain for other small countries. This clearly demonstrates the expansionist ideology. China (CCP) can gather data from these countries by exporting AI, embodying the so-called ‘independent controllability’ of China.”
Song Yangli revealed that in recent years, the CCP has issued management regulations for recommendation algorithms, deep synthesis, and generative AI services, mandating registration, security assessment, and content management for AI products offering services to the Chinese public. “Essentially, the CCP intends to centrally control AI content on a large scale, indicating that the National Security system is further intervening in the AI industry and information governance.” He believes that related laws and administrative regulations will be successively promulgated in the next one or two years.