The growing hacking capabilities of advanced artificial intelligence models are prompting Chinese authorities to pay greater attention to the technology’s potential national security risks.
Concerns have increased following demonstrations of AI systems capable of carrying out sophisticated cyberattacks with limited human oversight. A recent test by programmers showed that an AI system could be used to create a cyberattack worm capable of compromising large numbers of accounts on a major Chinese messaging platform.
Cybersecurity risks linked to autonomous AI agents are becoming a major concern as the technology develops rapidly. Both China and the United States face similar challenges involving AI safety, although discussions around these risks are handled differently in the two countries.
The issue is also expected to feature in discussions between the two countries as they compete for leadership in artificial intelligence. Potential cooperation could include shared safety evaluations for AI models and warning systems for emerging threats, although technological competition could make such efforts difficult.
China is also concerned that national security arguments could be used to restrict its technological development. At the same time, authorities have been strengthening regulations covering the use of AI, including rules protecting individuals from the unauthorised creation of AI-generated likenesses.
Chinese AI policy has largely focused on addressing existing social and security risks rather than hypothetical future scenarios. However, rapid advances in AI capabilities, high-profile security incidents and increased automation are pushing AI-related risks higher on the national security agenda.
Public reactions in China have included both concern and support for stronger regulation. Some people view the discovery of vulnerabilities as an opportunity to identify weaknesses and improve cybersecurity, while others argue that businesses cannot be expected to impose sufficient restrictions on themselves.
Another major concern is the ability of highly capable AI systems to break complex passwords and security codes. Experts warn that if AI can compromise critical systems or national security credentials, both China and the United States could face significant vulnerabilities without adequate safeguards.
