Methodology — NIST AI Risk Management Framework (AI RMF 1.0)

Each country is scored against the seven Trustworthy-AI characteristics defined in NIST AI RMF 1.0 (Jan 2023). Scores are 0–100 where higher = more risk. Characteristics without a direct signal in this dataset are marked "not measured" and excluded from the country's overall score — an absence of evidence is reported transparently, not imputed. Characteristic scores are rolled up into the four NIST Core functions (Govern / Map / Measure / Manage) to highlight where in the AI lifecycle each country is weakest. See the crosswalk table at the bottom for the mapping from risk dimensions to NIST characteristics and functions.

Overall NIST AI RMF risk by country

NIST Core functions — country × function heatmap

Shows the average risk across the characteristics NIST assigns to each Core function. Darker cells indicate the function that needs most attention for that country.

NIST Trustworthy-AI characteristics — country × characteristic heatmap

Granular view across all seven characteristics. A missing cell means the dataset does not provide a signal for that characteristic (gap in Measure).

Per-country NIST profile (radar)

Use the dropdown to switch country.

Country scorecard

NIST crosswalk

How the existing dataset signals map to the NIST AI RMF.