01
20%
Policy & Regulatory Framework
Whether a national AI strategy or law exists, how comprehensive it is, and whether enforcement mechanisms are in place.
Methodology · Transparency · Version-Controlled
How the Atlas AI Governance Observatory measures readiness
The Atlas AI Governance Observatory uses a transparent, evidence-based methodology to measure AI governance maturity and institutional readiness across 75+ countries. This page documents the pillars, scoring approach, data sources, and limitations behind every score Atlas publishes.
What the Scores Mean
Two headline metrics, two different questions
Atlas reports two distinct scores for every country profile. They are calculated independently and should not be read as interchangeable.
Governance Score
A weighted composite (0–100) reflecting a country's current AI governance maturity: how developed its policy, institutions, and safeguards are today, based on the six pillars below.
Readiness Index
A weighted composite (0–100) focused on institutional and technical capacity to implement future AI regulation, independent of whether formal policy already exists.
The Framework
Six pillars behind every country score
Each pillar is scored 0–100 using several sub-indicators, then combined into the Governance Score using the weights below.
02
18%
Institutional Capacity
Presence of a dedicated AI or data protection authority, its staffing, budget, and technical expertise.
03
17%
Data Governance & Privacy
Existence and enforcement of data protection law, cross-border data transfer rules, and consent frameworks.
04
15%
Transparency & Accountability
Algorithmic transparency requirements, public consultation processes, and redress mechanisms for affected citizens.
05
15%
Public-Interest & Rights Safeguards
Human rights impact assessments, protections for vulnerable groups, and information-integrity measures.
06
15%
Innovation Ecosystem & Cooperation
R&D investment, participation in global AI governance bodies, and regional cooperation on AI policy.
Scoring Process
From raw evidence to a published score
1
Indicator-level research
Each pillar is broken into 5–6 sub-indicators. Researchers score every sub-indicator for each country on a 0–100 scale using documented evidence.
2
Weighted aggregation
Sub-indicator scores are averaged into a pillar score, then combined using the fixed pillar weights to produce the Governance Score. The Readiness Index applies a separate weighting that favors institutional and technical capacity.
3
Expert review
Draft scores are reviewed by an expert panel and, where available, cross-checked with local Coalition members before publication.
4
Publication with changelog
Each country profile is published with a changelog recording when it was last updated and what changed, so scores remain traceable over time.
Data Sources
Where the evidence comes from
Primary Legal Sources
Official government gazettes, enacted legislation, and published national AI strategies and policy documents.
International Cross-Reference
Data from international bodies such as the OECD AI Policy Observatory, UNESCO, and the ITU, used to validate findings.
Coalition & Local Input
Civil society, academic, and policy-professional input from Atlas Global South AI Policy Coalition members on the ground.
Annual Desk Research
A structured annual review cycle plus an independent expert review panel that validates scores before publication.
Version Control
Every score is traceable and versioned
Each country score carries its own changelog recording when it was updated and why. The methodology itself is reviewed yearly; any material change increases its version number (for example, v1.0 to v1.1), so past scores can always be understood in the context of the methodology version used to produce them.
In the Interest of Transparency
Limitations of this methodology
Atlas AI Institute publishes these limitations alongside its scores so readers can interpret them appropriately:
- Scores reflect publicly available evidence at the time of research and may not capture unpublished or informal governance practices.
- Indicator weighting involves expert judgment; reasonable analysts could weight pillars differently.
- Rapidly changing regulatory environments may mean a country's actual status has shifted since its last review cycle.
- The Governance Score and Readiness Index measure different things and are not designed to be averaged into a single number.
- Atlas is an independent research institute; its scores are Atlas's own analysis and do not represent an official government or intergovernmental position.
Frequently Asked
Questions about the methodology
What is the difference between the Governance Score and the Readiness Index?
The Governance Score measures a country's current AI governance maturity across all six pillars. The Readiness Index measures institutional and technical capacity to implement future AI regulation. They are calculated separately and answer different questions.
How often are scores updated?
Atlas conducts a structured annual desk-research cycle, with expert review before publication. Each country profile shows a changelog of when it was last updated.
Where does Atlas get its data?
Primary legal sources such as government gazettes and enacted legislation, cross-referenced with international bodies like the OECD AI Policy Observatory, UNESCO, and the ITU, plus input from Atlas Global South AI Policy Coalition members.
Can the methodology change over time?
Yes. The methodology is reviewed yearly, and material changes increase its version number, so historical scores can be understood in the context of the methodology version used to produce them.
Explore the Data
See how countries score using this methodology
Explore the Atlas AI Governance Observatory for country-level Governance Scores, Readiness Index values, and full profiles across 75+ countries.