Can AI Decode How China Plans to Win the Future?
If China ends up leading the 21st-century economy, a large part of the credit will go to its industrial policy—a powerful blend of state-led planning, local experimentation, and bold investment in emerging technologies. While outsiders may view China’s strategy with both admiration and concern, a new study shows that artificial intelligence (AI) might now be the key to understanding how Beijing is quietly reshaping its industrial base.
In a groundbreaking collaboration, economists Hanming Fang (University of Pennsylvania), Ming Li and Guangli Lu (Chinese University of Hong Kong, Shenzhen) used Google’s Gemini, a large language model, to analyze millions of Chinese policy documents issued between 2000 and 2022. Their goal: map the full extent of China’s industrial policy landscape—one that extends far beyond high-profile campaigns like Made in China 2025.
A Labyrinth of Local Policies—Now Readable by AI
China publishes over 100,000 policy documents every year, many of them written by local governments with cryptic titles and vague language. Roughly one in five includes some form of industrial policy. This massive volume makes traditional analysis nearly impossible. That’s where AI comes in.
With Gemini’s help, the research team was able to identify and classify these documents, extracting detailed insights into which sectors were targeted, what tools were used, and how policies evolved over time. To ensure accuracy, they prompted the AI as an expert in Chinese industrial policy, ran spot checks, and even validated results using a second AI model from OpenAI.
Beyond Subsidies: A Broader Policy Toolkit
One of the study’s biggest takeaways is that industrial policy in China is far more diverse than commonly assumed. While subsidies appear in about 41% of policies, many other strategies are at play:
- Regulatory standards were used in 40% of central government policies.
- Consumer-side incentives (e.g., government purchases) doubled in popularity since 2000.
- Land and cheap credit appeared in less than 15% of cases.
- Explicit protectionism was mentioned in only 9% of documents.
Notably, only 29% of all policies targeted manufacturing. Increasingly, the focus has shifted to services, agriculture, and cutting-edge tech sectors like AI.
Imitation vs. Innovation: The Risk of Policy Convergence
The study also warns of an emerging risk: industrial isomorphism. Cities across the Yangtze River Delta, for instance, are all prioritizing “high-end manufacturing” and “new-generation information technology.” This kind of policy duplication, likely driven by pressure to align with national plans, can dilute the impact of industrial interventions.
Policies that work in early-adopter cities like Hangzhou or Shenzhen often fail when copied by less-developed municipalities. The “Manhattan Project” effect doesn’t scale easily—not every city can foster a tech revolution, no matter how bold the policy.
Mixed Results, but a Clear Direction
Despite the impressive volume and ambition of China’s efforts, the authors acknowledge that the link between industrial policy and firm productivity remains “mixed and tenuous.” Still, the use of AI opens a new window into policymaking itself—revealing patterns, inconsistencies, and opportunities for smarter interventions.
As AI becomes more deeply integrated into economic planning, China may start using these same tools to optimize future policy design. In a strategic twist, AI might not just be the target of policy—but also its architect.
Sources:
- The Economist. “How might China win the future? Ask Google’s AI.” (2024)
- Fang, Li & Lu. “Mapping Chinese Industrial Policy with LLMs.” Working Paper, 2024.
- U.S. Trade Representative. “2024 Report on Foreign Trade Barriers.”
#ChinaEconomy #ArtificialIntelligence #IndustrialPolicy #EconomicStrategy
#AIandGovernance
