The US scores highest on GIRAI's AI bias and discrimination measure at 90.8; 59 of 138 countries score zero

What the chart shows

GIRAI’s bias and unfair discrimination score assesses whether countries have laws, oversight bodies and enforcement to prevent discriminatory outcomes from AI, and whether civil society monitors bias independently. The United States scores highest at 90.8 out of 100, followed by Canada (85.7), New Zealand and the Netherlands (both 84.5) and Australia (83.5).

Most countries score low: 87 of the 138 score below 20 and 59 score zero. Only 19 score 50 or more.

Key numbers

How to read it

• GIRAI scores 138 countries from 0 to 100, based on an expert survey completed by in-country researchers between November 2023 and February 2024.
• This measures national laws, oversight and civil society activity, not how AI models themselves behave.
• The AI Index cautions that a low score does not necessarily mean a country disregards the issue; it can reflect an earlier stage of AI adoption or limited capacity to formalise AI-specific frameworks.
• The report shows this as a world map; we show the top 20 countries, and the CSV has all 138.