China’s domestic rules may make it easier to gather data and secure IP for AI development, but U.S. export controls and outbound investment limits give Washington the upper hand in denying critical hardware and models—shaping an exclusionary global AI race.

Main Finding

The chapter argues that divergent U.S. and Chinese legal and policy architectures—especially on data privacy, intellectual property, and export controls—shape each country’s comparative strengths in the AI race. China’s regulatory environment and state‑led coordination can improve access to training data and centralized support, while the United States retains advantages in private capital, LLM development, and the use of export controls to restrict critical inputs. Rising techno‑geopolitical competition is producing exclusionary rules that threaten “inclusive innovation” and global access to AI benefits.

Key Points

• Strategic context

  • AI is framed as a U.S.–China contest; U.S. firms dominate many leading LLMs (OpenAI, Google, Anthropic, Meta) but China’s AI industry has grown rapidly (large number of firms, fast growth in core AI industry and large model markets).
  • Governments on both sides play active roles: U.S. has evolving national AI strategies since 2016; China has an explicit state roadmap (Next‑Generation AI Development Plan) and local experimentation (ethics committees, industrial parks).

• Comparative indicators (2023 snapshot)

  • U.S.: much larger private investment in AI, more newly funded AI startups, more notable AI models, more top AI hubs.
  • China: a higher share of top AI publications in some measures, large number of AI enterprises, rapidly growing market value for large models.

• Three legal domains examined (chapter focus)

• Data privacy - AI requires massive datasets; use of personal data increases privacy and de‑anonymization risks. - United States: fragmented, sectoral, and state‑level privacy regime (e.g., CPRA in California, COPPA, HIPAA), no comprehensive federal GDPR‑style law as of end‑2024; FTC enforcement can address privacy abuses through consumer protection/antitrust authorities. - China: central and local measures (interim regulations, provision frameworks) and local governance (ethics committees, industrial parks) that facilitate lawful exchange/use of training data; state guidance is prominent. - Trade‑off emphasized: tighter privacy rules may constrain training data availability, while looser rules may ease model training but raise privacy and surveillance concerns.
• Intellectual property (IP) - (The chapter frames IP as a key infrastructure affecting incentives for model development and data use; specifics in chapter include how IP regimes interact with access to training data and commercialization incentives.)

• Export controls and capital/personnel restrictions - U.S. export controls on semiconductors, advanced chips, models, and outbound investments are a major lever to constrain Chinese progress. - China’s policies emphasize state coordination and targeted public investment to catch up and scale domestic capabilities.

• Normative and political dynamic

  • Both countries are increasingly prioritizing national security and competitive advantage over open, global flows of technology, capital, data, and people—producing exclusionary regulation.
  • This dynamic complicates achieving “inclusive innovation” (broad, equitable access to AI benefits).

Data & Methods

• Evidence and sources used in the chapter:
  • Quantitative indicators from secondary datasets and reports: Crunchbase (venture funding), China Center for Information Industry Development, Stanford AI Index, Global AI Innovation Index, Harvard Business Review analysis of AI hubs.
  • Benchmark comparisons of model performance reported in media (e.g., AIME math benchmark, Codeforces coding percentiles for Chinese models such as DeepSeek‑R1 vs. leading U.S. models).
  • Legal and policy documents: U.S. federal and state statutes and strategy documents (National AI R&D plans, executive orders, CPRA), Chinese national plans and interim AI regulations, local Chinese regulatory instruments.
  • Case examples and enforcement actions cited (FTC complaints and fines).
• Methodology:
  • Comparative legal and institutional analysis combining descriptive statistics, policy document review, and synthesis of scholarly and journalistic sources to infer how regulatory choices affect incentives, resource flows, and innovation outcomes.

Implications for AI Economics

• Data access and model performance

  • Regulatory differences in data governance materially affect the availability and cost of training data, which in turn influence model development trajectories and the competitive edge of firms/countries.
  • Economies with easier lawful access to large, diverse datasets may reduce training costs and accelerate model improvements, but at potential societal costs (privacy harms, surveillance externalities).

• Capital formation and innovation dynamics

  • The U.S. private‑venture ecosystem fuels rapid commercial innovation and many leading models; China’s state‑guided funding and industrial coordination can scale efforts quickly and align resources with national priorities.
  • Export controls and restrictions on cross‑border capital/personnel flows alter global innovation networks, raising the costs of specialization and potentially fragmenting markets.

• Market structure and competition policy

  • Fragmented U.S. privacy rules may both enable innovation (by avoiding blanket constraints) and create regulatory arbitrage between states and sectors; FTC enforcement and antitrust doctrine are being used to police abusive uses of AI.
  • Regulatory choices can create or entrench data incumbencies (data‑rich firms become harder to challenge), shaping market concentration and long‑run competition.

• Global public good and inclusive innovation

  • Exclusionary, security‑driven policies risk degrading global cooperation in AI R&D and limiting the diffusion of AI benefits, with distributional consequences across countries and within societies.
  • Policy choices that balance data protection, competition, and open scientific exchange will be critical for realizing more inclusive and socially beneficial AI-driven growth.

Note: This summary is based on the provided excerpt of the chapter (introduction, comparative overview, and the section on data privacy and U.S. privacy law). The chapter also addresses IP rights and export controls in detail; the implications section synthesizes those comparative claims as presented in the excerpt.