China’s Ministry of State Security (MSS) has issued its most detailed warning to date regarding artificial intelligence categorizing advanced AI systems as a severe threat to the nation’s political, ideological and critical infrastructure security.
Hostile foreign forces and cognitive warfare
In a public article written by Minister Chen Yixin, the spy agency accused external hostile forces of weaponizing generative AI technology to execute cheap, mass-scale disinformation campaigns.
The agency highlighted the proliferation of deepfake audio, video and synthetic text used to target Chinese public opinion framing the trend as an escalating attempt at “cognitive warfare” designed to undermine state stability.
Cyber espionage and model vulnerabilities
Beyond public opinion manipulation, the MSS pointed to concrete cyber vulnerabilities. The agency specifically flagged foreign intelligence actors for utilizing automated web crawlers and datamining to siphon sensitive trade secrets and private citizen data.
Furthermore, Western AI models – specifically noting high-tier systems like Anthropic’s Claude Mythos and OpenAI’s GPT-5.5-Cyber – were cited for advanced cyber capabilities that could potentially target China’s critical information networks.
Pushing for technological sovereignty
Rather than advocating for a slowdown in technological progress, Beijing emphasized the urgent necessity of achieving domestic technological sovereignty.
Minister Chen argued that China must secure full independent control over core capabilities particularly the production of high-end semiconductor chips required for training cutting-edge AI models to counter foreign dominance in the ongoing geopolitical race.
