BEIJING: China is increasingly concerned about the national security risks posed by rapidly advancing artificial intelligence, particularly as leading AI models demonstrate increasingly powerful hacking capabilities. Chinese policymakers have traditionally focused on immediate AI-related threats, including deep fakes and child safety. However, recent advances in AI-driven cyberattacks have prompted Beijing to take a closer look at the technology’s potential impact on national security. The head of China’s intelligence agency warned this week that AI could pose serious risks to national security. The concerns were further highlighted by a recent demonstration in California, where a small team of programmers claimed to have used AI to develop a cyberattack “worm” capable of potentially compromising millions of WeChat accounts. Kevin Xu, founder of US-based technology hedge fund Interconnected Capital, said China and the United States face more similarities than differences when it comes to AI safety concerns. He noted that such issues generally receive less public debate in China than in the United States. Experts say AI agents operating without human oversight could create significant cybersecurity risks, making them a major concern for governments. The growing threat is also increasing pressure on Chinese President Xi Jinping and US President Donald Trump to discuss ways of addressing AI-related risks. Sun Chenghao of Tsinghua University’s Centre for International Security and Strategy said both countries were developing a more realistic and shared understanding of AI risks, potentially creating greater urgency for cooperation. However, intense competition between Beijing and Washington could make cooperation difficult. Potential areas of collaboration include shared safety evaluations for AI models and warning systems to alert governments during emerging crises. China has already introduced regulations targeting some real-world risks associated with AI, including restrictions on creating AI-generated avatars based on a person’s likeness without their consent Experts warn that increasingly capable AI systems could eventually become highly effective at cracking complex passwords and sensitive security codes. Such developments, they say, could leave both China and the United States vulnerable if adequate cybersecurity defenses are not developed alongside advances in AI.
