December 20 2025

a collection of AI governance and china research ideas

core governance ideas in China

curated from conversations, essays and open literature. ai used to clean up clarify some points.
  • Unified Governance vs. Democracy: China's centralized system enables rapid, effective AI deployment and distribution compared to democratic processes.
  • Elite Politics and Technocrats: Increasing presence of CS/AI-educated leaders in the Central Committee; influence of overseas-returned experts and tech CEOs (e.g., profiles on Kai-Fu Lee, Ma Yi, Harry Shum).
  • Cultural Foundations of Big Tech: Origins and decision-making in Chinese companies shaped by national priorities.
  • Provincial Dynamics:
  • Deep competition between provinces.
  • Hangzhou/Zhejiang as innovation hub: Provincial government support for ecosystem; Alibaba's role; closer ties to Silicon Valley than Beijing.
  • Example: Wenfeng Liang (DeepSeek) as a "Demis Hassabis-like" figure in China.
  • Hot Take: China prioritizes practical adoption and scaling, potentially de-emphasizing existential/technical safety in favor of economic/utility focus.
  • Regulatory Processes: Licensing requirements for AI models emphasize controllability and alignment with state goals.
  • Pragmatic Materialism: Chinese AI discourse is utilitarian—focused on shipping products, revenue, and strategic objectives rather than superintelligence or existential debates. Emphasis on practical diffusion and adoption rates over frontier capabilities.
  • Topology of Chinese AI: Describes interconnected "circles" (government-backed institutions, enterprises, startups); highlights mission-driven younger founders (e.g., Liang Wenfeng of DeepSeek).
  • Vibe Shifts: Recent evangelism of ASI (Artificial Superintelligence) by figures like Alibaba's Eddie Wu signals potential departure from purely instrumentalist rhetoric.
  • Government-Industry Feedback: Continuous loops between policy and industry; confidence boosted by breakthroughs like DeepSeek.
  • Broader Context: China as both competitor and "obsessive ideal" for Silicon Valley; focus on infrastructure, scale, and socio-economic utility.

Hong Kong-Specific Notes

  • AI as Core Industry: Ambiguous—refers to production, application, or development? Analogous to declaring "energy" a core industry.
  • Public Sector Adoption: Plans to integrate AI into 200+ public services by 2027: https://www.scmp.com/news/hong-kong/society/article/3325910/hong-kong-roll-out-ai-use-200-public-service-procedures-end-2027

Implications for SMEs: Potential efficiency gains but challenges in adoption.

  • Related Resources:
    • Concordia AI report on Singapore AI Safety (relevant analogy for HK): https://concordia-ai.com/research/state-of-ai-safety-in-singapore/
    • HKCGI Thought Leadership: https://www.hkcgi.org.hk/thought-leadership/publication-detail/2620

ways to do AI research

  • Journalism/Connections: Reach out to HKU journalism (e.g., Karen Hao network) and SCMP reporters.
  • AI Safety Contacts: Leverage existing list of people in HK/Singapore working on AI safety.
  • Research Skills:
  • Follow Afra Wang's work: https://afraw.substack.com/
  • Use WeChat for primary sources on Chinese AI developments (and Baidu or the lesswrong/reddit equivalents of China)
  • the wechat ai field guide

general resources

  • Markus Anderljung's listhttps://www.matsprogram.org/mentors?track=governance-strategy
  • SPAR projects: https://sparai.org/projects/
  • Anton Leicht: https://www.antonleicht.me/
  • Concordia AI
  • Safe AI Forum: https://saif.org/

more references

  • Effective Altruism Forum China x AI Reference List (August 2025 update): https://forum.effectivealtruism.org/posts/jhuSbNinrrZ54s8Mw/china-x-ai-reference-list-august-2025-update
  • Google Doc on China AI Insights: https://docs.google.com/document/d/1OJcHhhBfNwEbeUaT-d4RIq58I1oJ3XGxu2yCzsnieuo/edit?tab=t.0#heading=h.9gc71zl8btmy

Note: Studying China is challenging due to rapid changes; even experts with language skills and lived experience struggle with up-to-date information. - Carnegie Endowment on China's AI Regulations and Decision-Making: https://carnegieendowment.org/research/2023/07/chinas-ai-regulations-and-how-they-get-made?lang=en - 80,000 Hours Career Review on China-Related AI Safety/Governance Paths: https://80000hours.org/career-reviews/china-related-ai-safety-and-governance-paths/?int_campaign=agi_lp