Beijing Tells Washington to Stop Hyping AI Dangers—But China’s Spy Chief Sees a Serious Threat at Home
The Paradox at the Heart of China’s AI Policy
A striking contradiction has emerged in China’s public posture toward artificial intelligence. On the international stage, Beijing has repeatedly urged Washington to abandon what it calls exaggerated warnings about the dangers of AI, arguing that alarmist narratives are little more than an attempt to contain China’s technological rise. Yet inside the country, the message is strikingly different. China’s own spy chief, Minister of State Security Chen Yixin, has framed artificial intelligence as a direct and growing threat to the security of the Communist Party—one that requires vigilance, defensive planning, and, at times, hard control. The two messages, delivered almost simultaneously, reveal a deep tension in how China understands the AI revolution. It wants the world to see the technology as a neutral, manageable tool. But its most senior security official is warning that AI, left unchecked, could be used to destabilize the very foundations of party rule. This is not simply a rhetorical mismatch. It is a window into the strategic thinking of a government that is trying to harness one of the most disruptive technologies in human history without losing control of the political order that depends on stability, surveillance, and centralized authority. For Beijing, the challenge is not whether to embrace artificial intelligence—it is how to do so without allowing the technology to become a vehicle for political change, social unrest, or foreign interference. And the answer, increasingly, is a policy of deliberate contradiction: reassuring the world that AI is safe while preparing the state’s security apparatus for the possibility that it is not.
A Public Message of Technological Optimism
For months, Chinese officials have used diplomatic forums, state media, and official statements to push back against the United States’ increasingly urgent warnings about artificial intelligence. When Washington introduced export controls on advanced semiconductors and raised concerns about AI in national-security contexts, Beijing responded with a familiar refrain: the United States is hyping the risks to justify technological protectionism. Chinese foreign ministry spokespersons have called on Washington to stop “ideologizing” technology and to avoid creating a new “digital Iron Curtain.” At the same time, China has promoted its own vision of AI governance—one that emphasizes safety, but also development, and that places the state at the center of oversight. In Beijing’s telling, AI is neither inherently good nor evil. It is a tool that can drive economic growth, modernize governance, and improve people’s lives, provided it is guided by the right principles. The danger, in this framing, lies not in the technology itself but in how it is used—and in the motives of those who seek to demonize it. This public posture is designed to reassure global investors, attract international partners, and position China as a responsible leader in a field that will define the twenty-first century. But it also serves a domestic purpose: projecting confidence and control to a population that is itself becoming increasingly aware of AI’s potential to disrupt jobs, privacy, and social order. By telling the United States not to exaggerate the risks, Beijing is also telling its own people that the party is in charge of the future. The subtext is clear: if AI is dangerous, it is dangerous only because of who might use it against China—not because of any fundamental flaw in the technology itself. And that is exactly the argument that China’s security establishment rejects.
A Much Darker Warning From China’s Intelligence Leadership
Behind the scenes, and sometimes in official settings, China’s security establishment has adopted a far more guarded tone. Chen Yixin, the minister of state security and a close ally of the party leadership, has described artificial intelligence as a major factor in what he calls “national security risks” facing the Communist Party. According to reports in state-backed outlets and commentary from security scholars, Chen has warned that AI can be weaponized by hostile forces to conduct espionage, spread disinformation, manipulate public opinion, and even incite social unrest. The language used in these warnings is significant. It speaks not just of cyberattacks or economic espionage, but of “subversion” and “infiltration”—the same vocabulary used to describe historical threats to party rule. In China’s national-security doctrine, the safety of the Communist Party is inseparable from the safety of the state. Any technology that enables independent organization, anonymous communication, or rapid dissemination of sensitive information is therefore treated as a potential threat. AI amplifies all of those possibilities. It can generate realistic fake media, target vulnerable populations with customized political messages, identify and track dissidents, and automate surveillance in ways that were unimaginable a decade ago. For the spy chief, these are not hypothetical risks. They are vulnerabilities that foreign intelligence agencies, terrorist groups, and domestic activists could exploit. Chen’s warnings are also linked to the party’s long-standing fear of “color revolutions”—the term it uses to describe any mass movement that could challenge Communist Party rule. In that context, AI is not just a tool for economic modernization. It is a potential weapon of political warfare, one that could be used to organize protests, spread rumors, or undermine public trust in the party. The more powerful AI becomes, the more dangerous it looks through the lens of regime security.
One Technology, Two Different Threats
The apparent contradiction between Beijing’s public diplomacy and its internal security messaging is not as puzzling as it first appears. In fact, it reflects a sophisticated—if uncomfortable—understanding of artificial intelligence as a dual-use technology. The same algorithms that power China’s facial-recognition systems, predictive policing, and social-credit initiatives can be turned against the state. The same large language models that generate official propaganda can be used to produce opposition narratives. The same data-driven platforms that enable digital authoritarianism can also provide the infrastructure for collective action. China’s leadership understands this dynamic all too well. That is why its approach to AI is defined by control as much as by innovation. The government has imposed strict regulations on algorithmic recommendation, deepfakes, and generative AI services, requiring companies to align their products with “core socialist values.” It has also invested heavily in its own AI capabilities, not only for economic development but for surveillance, censorship, and predictive analytics. From this perspective, the spy chief’s warning is not a rejection of AI. It is a call for stronger state intervention to ensure that AI serves the party’s interests rather than undermining them. The technology is seen as a strategic asset, but also as a potential liability. And the only way to manage that liability, in Beijing’s view, is to make sure the state controls the infrastructure, the data, and the political environment in which AI operates. That is why China has moved quickly to regulate generative AI, requiring companies to submit security assessments before releasing new services to the public. It is why the state maintains strict control over the data that AI systems train on. And it is why officials like Chen Yixin are so determined to place AI within the framework of national security—not because they fear the machines themselves, but because they fear what humans might do with them.
The Geopolitical Stakes of AI Governance
This dual approach has significant consequences for the global debate about AI safety and governance. The United States and its allies have increasingly focused on the existential and catastrophic risks of AI—the possibility that advanced systems could escape human control, enable bioweapons, or trigger a cyberwar. Beijing, by contrast, tends to frame the problem in terms of political security and regime stability. It is less concerned about AI’s hypothetical ability to wipe out humanity than about its practical ability to empower dissidents, expose corruption, or destabilize public trust. This difference in framing is not academic. It shapes policy. Washington has sought to impose export controls on cutting-edge chips, restrict Chinese access to American AI models, and build alliances around democratic values. China has responded by accelerating its drive for self-sufficiency, investing in domestic semiconductors, and courting countries in the Global South with an alternative model of AI governance. At international forums, Chinese representatives argue that AI governance should respect national sovereignty and allow each country to regulate according to its own conditions. That sounds reasonable. But in practice, it means something very specific: China wants the freedom to use AI as an instrument of state power while resisting international constraints on that use. At the same time, it wants to avoid a global arms race in AI rhetoric that could normalize the idea that AI is too dangerous to be left in the hands of authoritarian governments. The result is a growing disconnect in global AI governance. Western democracies are debating how to protect society from AI. China is debating how to protect the party from society. Both are talking about safety, but they mean very different things—and those differences are only becoming harder to bridge as the technology advances and the competition intensifies.
What the Contradiction Really Reveals
In the end, the contrast between Beijing’s public message and its spy chief’s warning is not a sign of confusion. It is a sign of strategic intent. The Chinese government wants to have it both ways. Abroad, it will continue to dismiss AI alarmism as a ploy by the United States to defend its technological dominance. At home, it will continue to treat artificial intelligence as a serious threat to Communist Party security—because, in a system built on centralized control, any technology that distributes power is a threat. The deeper truth is that Beijing’s leaders are not as confident as they appear. They know that AI is a revolutionary force that cannot be fully contained by censorship, regulation, or surveillance. They know that the same tools they use to monitor and shape public opinion can be copied, adapted, and turned against them. They know that the faster China advances in AI, the more exposed it becomes to the unpredictable consequences of the technology. That is why the spy chief’s voice is so important. He is not speaking out of turn. He is reminding the party that the greatest challenge of the AI era is not competing with the United States. It is maintaining control over a technology that, by its very nature, resists control. Beijing may urge Washington to calm down, but its own security apparatus is already preparing for the storm. The world would be wise to understand the difference between what China says about artificial intelligence and what it is actually doing to defend against it. Because when a government tells its allies that AI is safe, while telling its spies that AI is a danger to the state, the second message is usually the one that matters more.
