China's next big bet: invest in people for the AI era
China's most valuable growth lever isn't more factories or more capital. It's education that starts early, never stops, and keeps workers relevant as AI spreads through every industry.
That's the core message from Cai Fang, a leading labour economist at the Chinese Academy of Social Sciences. His take: policy must shift from building assets to building human capability-and AI should empower workers, not replace them.
Why this matters now
AI is moving into factories and offices at speed, turning job roles inside out. The question isn't whether work changes-it's whether people can keep up.
Global forecasts echo the scale of the shift. The World Economic Forum projected that structural changes would affect 22% of jobs between 2025 and 2030, with about 92 million roles displaced and 170 million created-a net gain of roughly 78 million positions worldwide. See the WEF report.
What Cai Fang is calling for
"As ageing accelerates, older workers' employment, skills and consumption will be critical to growth and social stability," Cai said in a recent discussion hosted by Tsinghua University's School of Public Policy and Management.
He argues for a decisive pivot: invest in early childhood education and lifelong learning; expand training for older workers so they can stay employable and even delay retirement; and build stronger social protection so transitions don't crush living standards.
Education, he added, can't be treated as a private advantage. It's a public necessity to help the next generation adapt to AI-driven disruption.
Signals from policymakers
Officials are acknowledging the shift. Zhang Yunming, vice-minister at the Ministry of Industry and Information Technology, said "restructuring does not mean disappearance, and iteration does not mean replacement."
He outlined an application-driven approach: use AI to upgrade workflows and production, raise workforce AI literacy, and grow cross-disciplinary talent fluent in both AI and manufacturing.
The Ministry of Human Resources and Social Security also said it will issue a document this year addressing AI's impact on employment-China's first systematic, central-level response to the issue.
Policy priorities that can't wait
- Early foundations: Expand quality early childhood programs, especially in under-served areas. Focus on language, numeracy, digital basics, and curiosity-skills that compound over a lifetime.
- Lifelong learning as infrastructure: Fund modular, stackable credentials; apprenticeship pathways; and training vouchers for workers and SMEs. Tie public financing to completion and employment outcomes.
- Older worker upskilling: Subsidize mid-career retraining, flexible scheduling, and on-the-job mentorships. Make digital literacy and AI co-pilot skills baseline for ages 45+.
- Stronger social protection: Modernize unemployment insurance, offer short-time work schemes, wage insurance for transitions, and portable benefits for gig and platform workers.
- Labour market systems: Upgrade job-matching platforms, recognize prior learning, and require skills-based hiring across public procurement.
- Public-interest floor for AI: Establish worker-impact assessments, transparent algorithmic management, and redress channels when automated systems affect pay, safety, or scheduling.
What this means for educators
- Redesign curricula around problem-solving, data literacy, and human-AI collaboration. Make project work the default.
- Partner with employers to keep course content current; co-develop micro-credentials that stack into degrees.
- Measure success by placement, wage growth, and skill usage on the job-not seat time.
- Open evening and weekend tracks for working adults, with childcare support where possible.
For program ideas and case studies, explore AI for Education.
What this means for government agencies
- Fund city and provincial reskilling hubs that pool public, university, and employer resources.
- Make training portable: one learner ID, unified transcripts, and national credential recognition.
- Link incentives to outcomes: higher subsidies for programs that place older and low-income workers into quality jobs.
- Publish regional AI job maps so schools and firms can target real demand.
- Set guardrails for AI in HR: bias testing, explainability for automated decisions, and minimum notice for algorithmic schedule changes.
For policy templates and procurement playbooks, see AI for Government.
What employers can do now
- Adopt AI co-pilots, then train teams to use them safely and productively. Productivity gains, shared fairly, fund more training.
- Move to skills-based hiring and internal mobility. Publish skill ladders so workers can plan their next step.
- Offer paid learning hours and recognized credentials. Tie promotions to verified skill growth.
- Include worker councils in AI rollouts to spot risks early and improve adoption.
How to track progress
- Training participation and completion, by age and region
- Time to re-employment after displacement
- Wage and productivity gains distributed across quintiles
- Share of 50+ workers in quality employment
- Percentage of roles using AI with documented risk controls
The bottom line
There's no automatic "trickle down" from AI productivity. Without strong institutions, gains pool at the top while workers absorb the shock.
China's advantage is speed and scale. If policy shifts from capital to people-early education, lifelong learning, and real protections-the labour market can absorb AI and come out stronger.
That's the test. Build capacity, not just capacity for output.
Your membership also unlocks:
