Updated
Updated · intelligence.org · Jul 30
China Renews Global AI Rules Push, Launches 2025 Safety Institute Over Existential Risks
Updated
Updated · intelligence.org · Jul 30

China Renews Global AI Rules Push, Launches 2025 Safety Institute Over Existential Risks

3 articles · Updated · intelligence.org · Jul 30

Summary

  • Xi Jinping used the 2026 World AI Forum to urge a consensus-based global AI governance framework, warning that fast-advancing AI could slip beyond human control without stronger safeguards.
  • China paired that call with institutional steps: its AI safety body, CnAISDA, was founded in June 2025 and says AI governance requires collective international participation.
  • Chinese officials have repeatedly framed the danger in stark terms, from Ding Xuexiang's warning against reckless competition to UN remarks calling for lifecycle monitoring, emergency response systems and a human "stop button."
  • That message extends to frontier-risk concerns: CnAISDA-linked scientists including Andrew Yao and Yi Zeng have warned about deception, catastrophic harm and even human extinction from superhuman AI.
  • China has pushed this line across multiple forums since at least 2023, including backing UN-led rulemaking, renewing its call in July 2025 and joining the 30-country Bletchley Declaration.

Insights

If Chinese leaders warn that humanity needs a stop button for AI, what concrete global enforcement would make that possible?
Can UN-led AI governance turn shared warnings from China and other powers into binding safeguards before frontier systems outpace regulation?
Is China’s push for global AI rules a real safety commitment—or a bid to shape who writes the world’s AI standards?