How the evidence is built
The atlas is assembled from public, citable datasets and rendered so every score can be traced back to its source and method.
Data & sources
The atlas joins public country datasets on ISO-3 codes. Missing sides stay blank, and only published dimension/pillar aggregates and URAI cases are used — full GIRAI indicator matrices are not joined in this release.
See researchGIRAI programme
Global Index on Responsible AI 2026
Country ranks and scores across five dimensions and three pillars, including the URAI penalty and recorded unacceptable-risk cases.
arXiv:2607.05404
The Jagged Global Economy
Structural national AI exposure, labour structure, gender gaps, remittance-adjusted exposure, occupation drivers, and adoption validation series (Murugan et al., CC BY 4.0).
AI Index
Stanford AI Index 2026
Microsoft GenAI diffusion (H1/H2 2025, Q1 2026), national AI strategy status, privacy/legislation status, and 2025 strategy mentions as policy context overlays.
How the atlas works
One consolidated, interactive view of governance, exposure, and adoption — designed for public scrutiny with methodology surfaced at every step.
MapLibre
One consolidated map
A single MapLibre vector choropleth on a CARTO basemap. Filters and an integrated country table sit in a left panel that collapses to a rail on select.
Deep panel
Country inspector
Clicking a country opens a right-hand data-viz inspector with gauges, a GIRAI radar, peer bars, and trend charts. Hover for rich per-country stats.
Transparency
Methodology popovers
Tap any “i” for per-index methodology — what it means, how it’s calculated, and its source — so every number can be traced and defended.