Annual global benchmark

AI Governance World Index 2026

A source-traceable assessment of the institutions, rules and operational capacity that make accountable AI possible.

No evidence, no score.

Missing evidence is not a zero. Documentary indicators are not imputed, company claims cannot substitute for public evidence, and Tier 4 media cannot be a primary scoring source.

Country results publish only after archived evidence, human review, the minimum completeness gate and full 195-country coverage for the annual edition.

Evidence collection status

0 verified observations across 0 countries. Rankings remain withheld until the publication gates are satisfied.

The ranking is in evidence review

All 195 country profiles are public. Annual scores will appear only after the complete, source-backed review and publication process.

Nine connected dimensions.

Each dimension contributes one ninth of the score; each contains five equally weighted indicators.

01

AI regulation maturity

Binding rules, institutional capacity, enforceability and demonstrated application of AI-specific governance.

11.12% weight
02

AI safety framework

Technical and institutional controls for identifying, testing, reporting and reducing AI risk.

11.11% weight
03

Transparency

Public visibility into AI use, documentation, procurement, oversight and participation.

11.11% weight
04

Public AI strategy

Currency, specificity, resourcing, accountability and evaluation of national AI strategy.

11.11% weight
05

Government adoption

Governance quality, operational capacity and evaluated outcomes of public-sector AI use.

11.11% weight
06

Privacy protection

Legal rights, independent supervision, automated-decision safeguards and enforcement of data protection.

11.11% weight
07

Research capability

Scale, specialization, impact, human capital and public infrastructure for AI research.

11.11% weight
08

AI investment

Transparent, sustained public and private investment in AI research, capability and infrastructure.

11.11% weight
09

Responsible AI policies

Operational protection of rights, oversight, contestability, inclusion and environmental responsibility.

11.11% weight

The method is part of the product.

Review every indicator, scoring anchor, source rule, missing-data constraint and known limitation before using the results.