Around the world, a growing awareness is emerging about the unrestrained development of artificial intelligence. Last week, the top leaders in the sector called for a slowdown in frontier research, whether it was Dario Amodei (Anthropic), Elon Musk (Grok), or Sam Altman, who even announced a pause on OpenAI’s IPO. In her State of the Union address, Ursula von der Leyen, the president of the European Commission, proposed a framework to govern the use of AI, at a moment when European sovereignty in this area seems more imperiled than ever.
Unlike their American or European counterparts, Chinese tech executives and regime officials have remained discreet about the AI rush. In China, the Communist Party’s concern centers on the potential consequences of this technology for its grip on the political process.
A major article was recently published by Chen Yixin , the Party secretary and minister of the State Security Department . Distributed on September 13 on the media China cyberespace , this text has been widely circulated and analyzed, both inside China and abroad.
The head of China’s intelligence services adopts all the trappings of official discourse in his address. He invokes the “important thought of Xi Jinping on building cyber power” and repeats the talking points used by the Chinese president at the World AI Conference (WAIC), held on July 17 in Shanghai.
Chen Yixin could have limited himself to arguing for global governance, especially since his piece appears ten days before Xi Jinping’s state visit to Washington, where Donald Trump is expected to propose a bilateral framework for dialogue on AI, the contours of which remain unclear. Instead, he focuses on domestic concerns and describes an “arena of battle” that has become “a new frontier of confrontation among great powers.”
The first of the six risks he lists is political. “Various hostile forces and individuals with malicious intent” would use generative AI to “fabricate political rumors, spread harmful information, provoke conflict, and inflame tensions in public opinion,” and to do so “on a large scale and at very low costs.” The minister draws a conclusion that calls for potentially drastic measures: “This situation threatens our political security, the security of our socialist system, and the security of our ideology.”
Next comes espionage. According to the head of China’s security services, foreign agencies “massively exploit” AI to collect “strategic national data, trade secrets, and personal information,” while naive Chinese users entrust their “sensitive information” to models designed outside the country. The admission is telling. Two days before the text’s publication, Anthropic accused seven Chinese labs, including Alibaba, DeepSeek and Moonshot, of having “distilled” Claude on an industrial scale through thousands of fake accounts and more than 150 million exchanges. Dependence on American models becomes, in Chen Yixin’s account, a vulnerability. Suspicion has hardened since Donald Trump authorized, on August 12, private digital pirates to conduct offensive operations in a cyberwar that the minister calls “industrialized and automated.”
The emphasis on ideology and domestic security must be taken literally. In the Marxist-Leninist framework that remains the Party’s, the algorithm is categorized among the means of production. It distributes labor (in China, Meituan and Ele.me dictate the exact timing of millions of deliveries), channels capital (the scores of Ant Group, the digital finance giant, have opened credit to hundreds of millions of borrowers outside the banking circuit), and daily manufactures public opinion (Douyin’s recommendation feed—TikTok’s equivalent—decides daily what more than 700 million Chinese see).
As early as 2019, data were elevated to the status of a production factor, and AI was classified among the “new-quality productive forces.” In Xi Jinping’s terms, social relations must “adapt to the requirements” of this rise, and it falls to the Party—“which leads everything”—to guide the adjustment before technology imposes it. In 2018, speaking at the National Cybersecurity Conference, Xi warned his cadres that the Party, if it “does not pass the Internet test, will not pass the test of maintaining power for the long term.”
Lenin once claimed to hold the “dominant heights” of the economy under the NEP; Xi’s Party intends to govern the algorithmic process just as it governs the economic process. Hence the first of Chen Yixin’s seven measures: to “maintain the Party’s leading role” and “the Party’s control over the Internet and data,” in order to “turn the Party’s leadership advantages into the governance efficiency of artificial intelligence.” The “healthy and orderly” development he envisions is designed to respond point by point to Xi Jinping’s warning about the “disorderly expansion of capital” issued at the end of 2020.
The precedent is etched in memory. In November 2020, the Party halted Ant Group’s IPO, and its founder, Jack Ma, vanished from public view for nearly three months. The ensuing antitrust campaign led to Alibaba’s record 18.2 billion yuan fine and brought all digital platforms to heel. No official signal suggests AI would meet the same treatment at this time. Chen Yixin’s text nonetheless reads like a warning to national champions as well as to foreign services. The Party encourages all technologies, so long as none threatens its power.
“Security, Xi Jinping said nearly a decade ago, is the precondition for development.”
Fully strengthen AI security guarantees, promote healthy and orderly development
Chen Yixin (陈一新) Minister of the State Security Department
Article published on September 13, 2026 in China cyberspace (中国网信 zhongguo wangxin)
We live in an era in which artificial intelligence is taking the lead in all technological revolutions and all shifts in industry. This sector has become the main battleground in international technological competition, and a new arena of confrontation among the world’s great powers. The Party Central Committee attaches great importance to the development of artificial intelligence. The Party’s general secretary, Xi Jinping, has spoken on this topic on numerous occasions. We must promote coordinated scientific and technological innovation, foster industrial development, and concentrate all our efforts on the enabling applications of AI. We must also strengthen supervisory and regulatory mechanisms for AI governance and reinforce its governance. We must study and thoroughly assimilate the spirit of Xi Jinping’s instructions, rigorously implement the overarching national security vision, accelerate the deployment of a risk-prevention and control system for AI, and promote the healthy and orderly development of this sector in our country.
The fundamental principles to be followed in AI governance
Since the 18th Party Congress in 2012, when Xi Jinping rose to the head of the Chinese Communist Party, he has repeatedly referenced artificial intelligence. He has stated, for example, that we must promote our country’s technological catch-up, upgrade our industry, and advance the productive forces. He has also asserted that China should progress in international technological competition and improve international governance in AI. To implement these directives, we must work in eight areas:
First, the strategic position of artificial intelligence. It must be clearly understood that AI is a strategic technology—it’s at the origin of new waves of technological revolutions and the transformation of industries today. In its wake, it can drive the development of all other sectors. Those who harness the economic opportunities AI offers will move with the times. This is why it should be treated as a strategic sector and placed at the heart of our national development program.
Second, we must pursue strategic policies to develop AI. We must make the most of the superiority of our national system and seize the opportunities presented by its renewal. We must persevere in self-sufficiency and give pride of place to “applied” AI. Innovation must become the primary engine of our economy, even as we remain mindful that security remains the most essential principle. We must also strengthen all of our country’s capabilities in AI.
Third, the question of AI values. It is important to remember that AI should contribute to the well-being of humanity as a whole. We must therefore promote healthy and orderly AI development in a direction that is beneficial, secure, and fair, never losing sight of the fact that the ultimate objective of scientific and technological innovation must be to meet the people’s aspirations for a better life. We should promote a close integration of AI with improvements in people’s living conditions, aiming to address fundamental human concerns such as education, health, elder care, and transportation.
Fourth, the governance of AI. We should coordinate development and security principles, strengthen our work in crafting and refining laws and regulations, public policy systems, and the design of enforcement mechanisms and ethical standards. We must establish a technical oversight platform, a risk-prevention and response system, and enhance analysis and prevention of potential risks associated with AI development, while protecting the interests of the people and national security.
Fifth, the core missions to which AI should be dedicated. We should deepen basic research, focus our efforts on mastering the most critical technologies such as semiconductors and software, and promote the integration of AI-driven innovation with industrial innovation to unlock new prospects for strategic industries and those of the future.
Sixth, we should promote the application of AI. We must fully implement the “AI Plus” initiative, ensure AI becomes the driver of a paradigm shift in scientific research, strengthen the integration of AI into industrial development, into our cultural identity, and into safeguarding people’s living conditions and social governance. We must lead in practical, industrial applications of AI and develop enabling applications across all sectors.
Seventh, the safeguards provided by AI. We should promote the construction of computing infrastructure, deepen data exploitation and use, openness, and data sharing. In this respect, we can leverage China’s data volumes and market scale. We should also advance policies on intellectual property, taxation, and public procurement, and improve mechanisms to support scientific research, the evaluation of capabilities, and the training of highly skilled workers.
Eighth, we should develop international cooperation in AI. AI should be a “global public good” and serve humanity as a whole. We must foster international cooperation to help Global South countries strengthen their technical capacities and reduce the “global digital divide.” We should encourage all parties to coordinate their development strategies, governance rules, and technical standards so as to establish, as quickly as possible, a framework for global governance and a consensus-based system of norms.
All of Xi Jinping’s interventions on AI are characterized by long-term vision, depth of thought, and the richness of their content. They function as fundamental guidelines, and it is important to fully grasp their essential principles. We must integrate them into every aspect of AI governance.
Fully recognizing the risks associated with artificial intelligence
We must also keep in mind that AI, precisely because it is the engine of new technological revolutions and industrial changes, affects all levels of human society. The diffusion of AI in our societies is rapid, deep, and covers a broad range of domains. AI is profoundly reshaping the structure of the global economy, challenging power dynamics and the paradigms of innovation and social governance. We can seize the opportunities it offers, but we must also clearly identify the risks it poses and keep our core principles in view.
First, AI has systemic implications for our country’s political and security environment. By June 2026, the number of AI users in China exceeded 500 million, with an average daily volume of 140 trillion “token calls.” AI-integrated applications have rapidly become ubiquitous. Various hostile forces and malicious actors now abuse generative AI technologies to produce fake audio or video recordings, “deepfakes,” AI-generated images and texts, and “weapons of trolls,” with the aim of inventing political rumors, spreading harmful information, provoking conflict, and stoking tensions in public opinion.
AI enables them to conduct these actions on a large scale and at very low costs. This situation threatens our political security, the security of our socialist system, and the security of our ideology.
Second, AI is increasingly used to conduct and prevent cyber-attacks. As evidenced by Claude Mythos 5 or the latest GPT-5.5-Cyber, recently launched by American tech firms, AI technologies have advanced substantially in detecting software vulnerabilities.
They can also develop malware and be used as weapons, while cyber-attacks can now be industrialized and automated. We are even witnessing battles between AIs. Some countries and organizations are now able to rapidly identify large numbers of security vulnerabilities, connect automatically to organize cyber-attacks, and carry out complex hacking missions.
The technical threshold to engage in such activities and the entry costs have fallen significantly, posing serious dangers to our country’s critical IT infrastructure.
Third, the risks of leaking confidential information are increasing dramatically. Foreign intelligence services massively exploit certain technologies such as intelligent index robots, smart data exploration, and intelligent image analysis to collect sensitive information at a large scale, including strategic national data, trade secrets, and personal data of Chinese citizens. Some Chinese users, not sufficiently aware of data security, use foreign AI models to process sensitive information and thus export data abroad, causing widespread data leaks from our territory. Open-source AI tools like “OpenClaw,” which saw explosive growth in the first half of 2026, also suffer from structural problems, notably the risk of remote taking control of device administration rights or leakage of sensitive information about users.
Fourth, the monopolistic concentration phenomenon intensifies imbalances in science and technology development. Some countries, leveraging advantages in fundamental AI theory, model architecture, and computing power, strictly limit access to the technology by employing measures such as entity lists, technology controls, monopoly of industrial standards, and closed ecosystems.
Using the pretext of national security, they restrict access for Global South countries to cutting-edge technological equipment, thereby limiting the strengthening of their own scientific and technological innovation capabilities. These restrictions fragment global AI supply and value chains.
Fifth, structural problems affect social governance systems. AI applications introduce many uncertainties into social governance and public order: the “black box” nature of algorithms and data poisoning could amplify social inequalities. Misuse of personal data in processing can erode user trust. Automated decision-making by systems also raises accountability issues. The rapid development of AI has, for example, outpaced traditional governance frameworks for technology. Legal norms, ethical principles, and governance mechanisms often lag behind AI, making it difficult to implement an effective system of institutional constraints and regulation.
Sixth, AI profoundly alters the nature of armed conflict. AI technologies push warfare into an “age of intelligence.” Whoever can achieve finer perception, more precise judgment, and more accurate strikes via algorithms will hold an advantage on the battlefield. This was evident in the US-Israel-Iran conflict, where military AI applications have given rise to intelligent intelligence integration, intelligent decision-support, intelligent target identification, and intelligent combat support. AI is shifting from a supporting role to a decisive one on the battlefield.
Measures to advance AI governance
faced with the risks linked to the rapid evolution of AI technologies, we must rely on the many significant expositions by Secretary-General Xi Jinping, implement the reform measures laid out in the 15th Five-Year Plan (2026-2030) to modernize the system and capabilities of national security, accelerate the improvement of AI governance, and effectively prevent and address security risks in this domain.
First, we must uphold the Party’s leading role. We should study and apply the “Xi Jinping Thought on Socialism with Chinese Characteristics for a New Era,” in particular Xi’s important ideas on building a digital power. We should also maintain the Party’s control over the Internet and data, integrate AI governance into the broader framework of building a digital power and a “Digital China,” improve governance structures, maintain the correct direction, and transform the Party’s leadership advantages into governance effectiveness for artificial intelligence. We must guide AI toward progress and virtue, ensuring its sustainable evolution.
Second, we must strengthen our system’s capacity for coordinated action. We should consolidate the foundations of our national security strategy, fully integrate AI security into security efforts in other strategic domains, harness the advantages of the new national system, establish mechanisms for cross-disciplinary and inter-ministerial collaboration, promote data pooling and sharing, continually optimize consultation, coordination, and decision-support mechanisms, risk assessment, and security analysis. We must also advance the coordinated improvement of legal, strategic, policy, and risk-prevention frameworks.
Third, we must bolster risk detection and prevention. We should establish and refine a platform for AI security oversight and regulation, deepen threat assessments and analyses, advance research on major questions such as AI development trends, scientifically analyze potential risks associated with emerging technologies, precisely identify relevant laws and developments, and proactively address vulnerabilities and potential risks. We should regularly publish risk-prevention guidelines to guide various actors, including industry companies and government services.
Fourth, intensify efforts in scientific and technological innovation. We must intensify research into the fundamental principles of AI and key technologies such as semiconductors, development frameworks, and application software; accelerate the construction of AI infrastructure, including national and local data pools and the East-West Data Transfer Project network; encourage and direct different actors (the government, enterprises, and research institutes) to strengthen collaboration; promote synergy between industry, higher education, and research to build an AI ecosystem; guarantee autonomy and control of key technologies; firmly safeguard technological sovereignty; and achieve positive interaction and coordinated progress between innovation and governance in security.
Fifth, perfect the legal safeguards system. We must fully implement the “Cybersecurity Law of the People’s Republic of China” and the “Counter-Espionage Law,” accelerate the adoption of specific laws and regulations related to AI R&D, deployment, and governance; in matters such as algorithm security, data protection, ethical standards, and privacy, we must establish a comprehensive set of rules and regulatory norms spanning the entire chain—from research and development to implementation, risk prevention and control, and accountability. We should strengthen oversight mechanisms and enforcement capabilities, reinforce the legal responsibilities of regulators and AI-focused enterprises, and sanction illegal and criminal activities such as data theft or the creation and dissemination of harmful information via AI under the law.
Sixth, innovate in propaganda and mobilization methods. By leveraging important events such as National Security Education Day on April 15 and Cybersecurity Awareness Week, it is appropriate to use both traditional media and new media platforms to broadly disseminate Xi Jinping’s important speeches on AI governance. We should also deepen policy explanation, legal interpretation, case analysis, and risk prevention, while strongly denouncing the actions of certain countries that erect technological barriers and use AI to infiltrate our country and steal sensitive information, and respond to the slander, denigration, repression, and blockade we may encounter.
Seventh, strengthen international cooperation. By adhering to the vision of global security governance based on consultation, joint construction, and shared benefits, we should deepen international science and technology cooperation, firmly oppose hegemonism, technical barriers, and fragmentation, and establish an international system of fair and open AI rules. We should promote technological exchanges and openness and sharing of results, foster inclusive and balanced development, reduce the digital divide, and ensure that AI benefits the people and the world at large.