The 15th Five-Year Plan Outline, covering 2026–2030, describes its role as clarifying national strategic intentions and defining government priorities. It is not the same as a law. Not every sentence creates an immediate legal obligation for every company or public institution, but it influence shapes the policies that follow later.

The policy process usually moves through several layers:

→ Five-Year Plan → specialised and sectoral plans → laws, regulations and standards → ministerial and local implementation → projects, funding and institutional practice

The national plan identifies the direction. Later measures determine who must act, what procedures apply and how implementation will be assessed.

The earlier stage focused heavily on building the technology: computing capacity, models, algorithms, data resources, research capability and AI companies. The next challenge is broader: how to integrate AI into the economy and society at scale.

The plan identifies four connected tasks.

First, China needs stronger technical foundations, including computing infrastructure, core algorithms, high-quality datasets and supporting software and hardware.

Second, it wants to move AI from isolated applications into manufacturing, scientific research, education, healthcare, consumer services and public administration.

Third, wider adoption will change the structure of work. Some tasks will be accelerated, some occupations will expand, and new forms of human–AI collaboration will emerge.

Fourth, AI systems create questions of safety, transparency, responsibility and rights that cannot be addressed through technical development alone.

The underlying policy problem is therefore not a shortage of AI applications by itself. It is the gap between the speed of technological change and the ability of institutions to adapt around it.

China’s response combines two forms of investment:investment in technology, infrastructure and industry; investment in people, public systems and governance capacity.

This is visible in the language of the Employment-First Strategy Plan for the 15th Five-Year Plan Period, which explicitly calls for investment in physical development and investment in people to be considered together.

The plan’s digital and intelligent development agenda prioritises computing infrastructure, model and algorithm development, high-quality datasets, industrial software and wider adoption across manufacturing, education, healthcare, employment, consumption and government. These priorities are set out in the official overview of the plan’s digital and intelligent development chapter.

The broader direction was already articulated in the State Council’s Opinion on Further Implementing the “AI Plus” Initiative. That document describes a transition toward an intelligent economy and intelligent society shaped by human–AI collaboration, cross-sector integration and new forms of production and employment.

The plan calls for the safe and orderly deployment of large AI models in government. It refers to public-service systems that can identify needs more precisely, plan services proactively, support end-to-end administrative processing and strengthen risk detection and response.

This implies a possible movement from primarily reactive administration toward more integrated and anticipatory public services. It should not be read as transferring public authority entirely to automated systems. The stated direction remains regulated deployment with institutional responsibility and human oversight.

The plan also calls for stronger AI laws, policy rules, ethical guidelines, algorithm filing, transparency management, safety assessment and responsibility allocation.

It proposes lifecycle risk management covering development, testing, deployment, monitoring and incident response. The direction is clear, but the boundaries are not yet complete. A central unresolved issue is how responsibility will be divided when an AI model is developed by one organisation, supplied by another, deployed by a third and used by individual employees.

The Employment-First Strategy Plan for the 15th Five-Year Plan Period calls for employment impact to be considered in major policies, projects and productive-force allocation. It also requires monitoring of labour demand, job creation and reduction, and unemployment risk.

This does not mean that every organisation deploying AI must already conduct a formal employment-impact assessment. It does show that employment consequences are moving closer to the beginning of economic and technology policymaking.

The employment plan calls for exploration of new forms of human–AI collaborative work. An occupation may remain while research, drafting, scheduling, analysis, customer service or quality control are redistributed between people and AI systems.

The employment plan promotes a model combining job demand, skills training, skills assessment and employment services. It gives enterprises and industry-led programmes a central role in vocational and transition training.

The plan includes large-scale youth skills and employment-placement programmes, more technical and managerial placements, work experience and enterprise-led development for new employees.

If AI compresses entry-level work, what will replace the learning that entry-level work once provided?

The plan does not fully answer this question, but its emphasis on work-based learning, placements and enterprise responsibility suggests that youth employment is being treated as a capability-development issue, not only a question of access to a first job.

The employment plan calls on platform companies to regulate algorithms, improve transparency and protect workers’ rights to know about, participate in and make choices concerning algorithmic rules.

An official explanation of the employment plan is available from the National Development and Reform Commission and the Ministry of Human Resources and Social Security.

Implementation of “AI Plus”: sectoral action plans, demonstration projects, funding programmes and local initiatives.

Sector-specific governance: different expectations for manufacturing, healthcare, finance, education, employment and public administration.

Data and algorithm governance: data quality, algorithm filing, transparency and platform rules.

Safety assessment and lifecycle management: testing, documentation, monitoring, incident reporting and responsibility.

Public-sector and state-linked implementation: government departments, public institutions, state-owned enterprises and publicly supported pilots.

Workforce development: occupational standards, enterprise-led training and worker-transition programmes.

Employment monitoring and algorithmic work: risk-warning systems and the governance of automated management.

Regional variation: different speeds and implementation models across provinces, cities and sectors.

These are areas of interpretation and observation. They are not predictions of specific legal outcomes or operational recommendations.