Zeng et al. 2025 — AGILE Index 2025
Citation
Zeng, Y., Lu, E., Guo, X., Huangfu, C., Xie, J., Chen, Y., Wang, Z., Liang, D., Cao, G., Wang, J., Ruan, Z., Guan, X., & Younas, A. (2025). AI Governance InternationaL Evaluation Index (AGILE Index) 2025. arXiv:2507.11546.
Summary
Comprehensive framework for evaluating national AI governance maturity across 40 countries. Builds on 2024 pilot (14 countries) with expanded scope and refined methodology. Uses 4 pillars, 17 dimensions, and 43 indicators to assess governance capacity and effectiveness.
Core Contribution
Operationalizes the principle “the level of governance should match the level of development” into a measurable index. Distinguishes between governance capacity (instruments, environment) and governance effectiveness (actual outcomes, public trust).
Key Findings
- China overtook US in rankings due to more consistent AI policy; US dropped after revoking Executive Order 14110
- AI incidents doubled in 2024
- China and US account for ~70% of open AI models and datasets
- 14% of AI papers now focus on governance topics
- High-income countries lead in development and instruments; lower-income countries perform better in public trust and acceptance
- Clear positive correlation between index scores and per capita GDP
Framing Value
The governance-development proportionality principle is the transferable insight. Not that governance should constrain development, but that governance sophistication should scale with technological capability. This applies beyond national policy to organizational AI governance, product development, and risk management.
Extractions
- 05-atom—governance-development-matching-principle
- 05-atom—development-governance-gap-pattern
- 05-atom—income-governance-paradox
- 05-molecule—agile-index-framework
- 05-molecule—country-typology-ai-governance