Beijing's AI Alarm: Spy Agency Flags Cyber Hacking, 'Cognitive Warfare' as Top National Security Threats

China's top spy agency, the Ministry of State Security (MSS), has issued its most detailed public warning yet about the serious national security risks posed by artificial intelligence (AI), especially its hacking abilities. This unprecedented move, articulated by MSS head Chen Yixin in mid-September 2026, marks a significant shift in Beijing focus, moving beyond common concerns like deepfakes and child safety to openly address external cyber threats and their potential to destabilize the nation. The core of Beijing concern lies in what Chen described as 'cognitive warfare' and highly efficient cyberattacks. He warned that 'hostile forces' could use generative AI to create deepfakes and automated comment armies to spread political rumors, harmful information, and incite division, directly threatening China's political and ideological stability. The MSS chief also highlighted how AI could lower the bar for sophisticated cyberattacks on critical infrastructure, even specifically naming U.S. AI models like Anthropic Claude Mythos and OpenAI GPT-5.5-Cyber as examples of this 'disruptive upgrade' in hacking capabilities. This comes after reports of a U.S.-based team using AI to develop a 'worm' capable of compromising millions of WeChat accounts, adding to the urgency. Simultaneously, U.S. agencies have warned that China-based AI companies are allegedly using 'knowledge distillation' to extract proprietary models from leading American AI firms, intensifying the technological rivalry. Looking ahead, these stark warnings set the stage for crucial discussions between Chinese President Xi Jinping and U.S. President Donald Trump, who are expected to address AI governance at upcoming talks on September 24. While both global powers acknowledge the escalating risks of AI, fierce competition remains a hurdle for genuine cooperation on safety measures. Beijing, through its Cyberspace Administration of China, already has various AI regulations in place, treating AI as a manageable technology akin to electricity, but the recent 'warning shots' are pushing national security to the top of its regulatory agenda, suggesting even stricter oversight may be on the horizon.