Measuring how AI governance frameworks are built
Most comparisons of international AI governance ask what a framework says. This one asks how it is built: who holds the authority, how hard it is to change once agreed, and what it costs to put into practice.
What the index does
The AI Governance Design Index scores international AI governance frameworks on three structural axes — power concentration, reversibility and efficiency — each built from four equally weighted features with published scoring anchors.
The scores are descriptive, not evaluative. A framework that concentrates authority is not thereby better or worse than one that does not; the point is to make the design trade-offs visible and comparable, so that arguments about which trade-offs are worth making can start from the same picture.
Read the full index and scoring rules, go straight to the framework scores, or see what the index cannot do.
Who we are
-
Camilla Balbis
AI Governance & Safety Policy · Research Fellow, SPAR · Researcher, AIGS Canada · ISO/IEC 42001 Lead Auditor
Camilla Balbis works at the intersection of AI governance and security. She is a SPAR Research Fellow at TAICI, where she helped develop ScamBench, a benchmark for evaluating AI-enabled scams, and a Policy Researcher at AIGS Canada, where she researches the country's domestic preparedness for advanced AI systems. She has judged and presented on AI safety work including the Global South Hackathon 2026, where she spoke on "Who Gets to Govern AI?", and the Public Health SPOTlight discussion on safe AI integration in healthcare. She is co-founder of Kosmiai, an AI governance advisory helping SMEs and nonprofits adopt AI safely, and an ISO/IEC 42001 Lead Auditor with deep expertise in the EU AI Act.
-
Parveen Sheikh
She/Her · Founder, NISA · Gender Equity, Livelihoods & Social Inclusion · Community Empowerment · Intersectional Feminist · Cinephile · Vegetarian
Bio to come.
-
Jeanne MJ Vincendeau
Political Violence & AI Governance · Double MA IR, LUISS × CFAU · Co-Founder, Horizon AGI · U4SSC Working Group 1
Bio to come.
Independence and funding
This project is independent and unfunded. It is not produced by, for, or on behalf of any of the institutions the authors are otherwise affiliated with, and none of the bodies scored in the index were consulted on their scores.
Balbis, C., Sheikh, P. and Vincendeau, J. M. (2026). AI Governance Design Index, version 2.0. Available at https://example.org/