China Plans Broad AI Guardrails After Global Security Lapses

China is accelerating legislation to regulate artificial intelligence across its full life cycle, from the underlying data, computing capacity and algorithms to specific applications. The proposed framework would set safety standards and ethical guidelines, with regular reviews so outdated rules can be revised. It comes as Beijing seeks to strengthen its AI industry against US competitors while responding to fears that powerful systems could escape control or enable more damaging cyberattacks. Recent security incidents and warnings from technology leaders have sharpened those concerns. The rules will affect leading Chinese developers, including DeepSeek, Moonshot and Alibaba.

Beijing says it will temper compliance demands with a transition period, practical guidance and sector-specific regulatory sandboxes where companies can test systems with more flexibility. Developers and operators will be expected to manage risks and disclose safety incidents and vulnerabilities. Authorities will also track AI’s effects on employment and provide protections for workers displaced by the technology. At the same time, China plans to continue encouraging open AI development through 2030, reflecting its effort to balance safety oversight with rapid innovation. Officials are also calling for international cooperation on safety governance, vulnerability disclosure and responses to risks, arguing that no company or country can manage AI threats alone. China and the United States have already agreed to seek a shared understanding of AI’s benefits and dangers.