China names US AI models as cybersecurity threat

China's Minister of State Security, Chen Yixin, has named two advanced U.S. artificial intelligence models as specific cybersecurity threats to the nation's critical infrastructure. In a signed article published on September 13, 2026, in the official journal China Cyberspace, Chen identified Anthropic's Claude Mythos and OpenAI's GPT5.5Cyber as examples of technology that is markedly raising the efficiency and weaponization of vulnerability discovery and malware development. The article, which represents one of Beijing's most detailed public warnings on AI security, outlines six primary risks posed by the technology. Chen placed the systemic threat to political security at the top of the list, stating that hostile forces are abusing generative AI to fabricate political rumors and wage cognitive warfare. The second risk, where the U.S. models are named, describes a disruptive upgrade in cyber offense and defense, characterized by industrialized vulnerabilities and fully automated attacks. Chen's assessment contrasts with the dominant AI safety discourse in the United States, which has focused on existential risks and the potential for AI to slip out of human control.