China’s Minister of State Security, Chen Yixin, warned that artificial intelligence could threaten the Communist Party’s control and the country’s political security, calling for stronger party oversight and tighter government regulation of the technology to counter risks including disinformation, espionage, cyberattacks and military competition.
Chen Yixin calls for tighter government oversight
Chen set out his position in an article published on Sunday by the state-run magazine China Cyberspace. He said tighter oversight was necessary to protect domestic political stability and counter disinformation campaigns and cyberattacks he attributed to “hostile forces”, as well as to address military competition with countries such as the United States.
The Chinese official cited US technology companies and their artificial intelligence models, including “Claude Mythos” and “GPT-5.5-Cyber”, as examples of new cybersecurity threats. He also warned that foreign intelligence agencies could use AI in extensive espionage operations and attacks on China’s critical infrastructure.
Henry Gao, a law professor at Singapore Management University, said the warning reflected China’s treatment of foreign AI models as a real threat. Beijing, he said, viewed AI safety through a competitive security lens, and data sovereignty and political security would not be up for negotiation in international safety agreements.
US-China talks precede Washington meeting
Kyle Chan, a researcher at the Brookings Institution, said authorities feared the spread of fabricated content attributed to Chinese officials and the resulting disruption. Chen also said AI had lowered the technical and financial barriers to carrying out cyberattacks, posing risks to China’s information infrastructure.
His warning came after a California research team disclosed a hacking tool capable of taking control of millions of users’ accounts on the WeChat app. The article also addressed the risk of information leaks when using foreign models.
Data leaks widen risks from foreign models
Anthropic had said that the Chinese startup Moonshot AI had routed its users’ queries to the Claude model, some of which included surveillance footage linked to the Chinese military and private information about Chinese technology companies. On the military front, Chen said the US military’s use of AI in the context of the US-Israeli confrontation with Iran provided an example of how the nature of warfare was changing.
He added that possessing algorithms that improve sensing and targeting accuracy gives the party using them the initiative on the battlefield. China classifies AI as a strategic technology for development and competition with the United States, while also issuing regulatory warnings and guidance.
The Ministry of Industry and Information Technology announced in July that it had identified security risks in some versions of the “Claude Code” tool. Chinese regulators also warned of the risks posed by AI-powered assistants such as “OpenClaw”.
Beijing rejects threat and malicious competition narratives
Meanwhile, Chinese Foreign Ministry spokesman Guo Jiakun on Monday rejected what he called threat and malicious competition narratives about China’s AI sector, saying they made international cooperation on regulating the technology more difficult. He was responding to Western calls for tougher restrictions on Chinese companies.
Anthropic CEO Dario Amodei had called in an open letter for democratic countries to maintain their lead over China and backed restricting sales of advanced AI chips to the country. The Communist Party-affiliated Global Times described his call to contain China’s development as an “unrealistic illusion”.
