China's NDRC Lays Out a 15th Five-Year AI Governance Framework

On November 3, China's National Development and Reform Commission (NDRC) held a press conference on cultivating new growth drivers and responded to questions on balancing AI safety and development. Vice Director Wang Ruomeng outlined a four-pillar governance plan for the 15th Five-Year Plan period, including a foundational comprehensive AI law and AI regulatory sandboxes.

Four Governance Pillars

The NDRC said it will always balance development and security, embedding safety across every stage of AI research, deployment, and governance.

1. Build a sound governance rule system and define compliance boundaries — accelerate foundational comprehensive AI legislation and establish lifecycle safety standards and ethics guidelines.
2. Optimize the regulatory environment and enhance adaptability — set transitional periods, provide compliance guidance, and build AI regulatory sandboxes with clear application, testing, use, and exit mechanisms.
3. Clarify multi-subject collaborative governance responsibilities across government, industry bodies, enterprises, and the public.
4. Actively respond to AI's social impact, ensuring technology stays people-centered and beneficial.

Legislative and Regulatory Moves

Concrete steps were laid out for the planning window.

• The NDRC will systematize China's AI governance experience and push forward a foundational comprehensive AI law to form the backbone of an AI legal framework.
• It will improve legislation covering data, compute, algorithms, and key application scenarios, with periodic evaluation and dynamic adjustment as the technology iterates.
• Regulatory sandboxes will be built for different domains with pre-set risk buffers and wider room for trial and error.

Employment and New Occupations

The commission addressed the labor-market effects of AI directly.

• The NDRC will closely monitor AI's impact on employment structure and push new roles such as model trainers and data annotators into the national occupational classification system.
• For job-displacement risk, it will strengthen monitoring and early warning and align the social-security system with changing work patterns.

Key Data Points

• 4 governance pillars announced
• 1 foundational comprehensive AI law in drafting
• 2 new AI occupations to be classified (model trainers, data annotators)
• 15th Five-Year Plan window
• press conference date November 3, 2026

Sources: China News Service, Southern Metropolis Daily, NetEase, threerouter