New data shows artificial intelligence is spreading faster than governments can regulate it, raising concerns that people’s rights may be left unprotected.
Global Index reveals uneven AI governance
The second edition of the Global Index on Responsible AI (GIRAI) compares 135 countries across five human‑rights‑linked dimensions, including AI use in public services, ethics, inclusion, labour and trust. The study, compiled by the Global Center on AI Governance, finds that while more nations are adopting AI strategies and laws, the average score remains low—about 35 out of 100—and implementation lags behind.
Roughly 53 % of the global population has used generative AI tools, yet only 55 % of the identified frameworks show any evidence of implementation. In the Global South, that figure drops to 45 %, highlighting a widening gap between policy ambition and practical action.
Related: Globant makes Fast Company’s brands to watch
“Responsible AI cannot be secured through principles alone,” said Rachel Adams, founder and CEO of the Global Center on AI Governance. “The second edition of GIRAI shows a persistent gap between responsible AI as a commitment and responsible AI as a capability.”
Key weaknesses in oversight and transparency
Gender‑related frameworks have increased, with 29 new countries adding gender considerations since the first edition. Still, only 24 of the 55 nations with such policies show evidence of implementation, and protection from gendered harms remains limited.
Environmental impact is another blind spot. Only 27 % of countries have frameworks that address AI’s energy or water use, and most of those are non‑binding. The index therefore highlights a lack of full oversight for AI’s ecological footprint.
Technical safeguards are expanding, but human harms are “under‑addressed,” suggesting a misalignment of priorities.
Related: Pearl Perfection: Classic and Elegant Engagement Rings for Women
In the African region, the average GIRAI score is 22, the lowest worldwide. Nigeria, Egypt, Kenya, Ghana, Benin and Morocco are the only African nations scoring above the global average. Even where policy coverage is broad, implementation rates hover around 20 %.
Average scores range from 55 in the Global North to 27 in the Global South. Only 54 % of countries have adopted a national AI policy or equivalent framework, and a mere 27 % have operational mechanisms for civil society participation.
AI literacy frameworks exist in 71 countries, and 53 % of nations have some policy on AI education. Yet, only 41 % address children’s rights in AI, and just 27 % of those show concrete implementation.
Related: Kano tackles water shortage with solar boreholes
Local‑language initiatives are gaining traction, with 39 % of countries pursuing government‑led projects to adapt AI for diverse linguistic contexts.
More than 68,000 data points were collected by local experts, providing a detailed view of where governments stand on AI governance.
Without a shared rights‑based floor, the report warns that “interoperability risks serving markets before it protects people.” The authors call for stronger institutions, independent oversight, public disclosure, monitoring systems and accessible routes for redress to bridge the divide between AI’s rapid expansion and the slower pace of protective regulation.
