A Comparative Framework for AI Regulatory Policy

Objectives

Over the past three years, the AI governance landscape has become increasingly complex, with distinct approaches to regulatory policy emerging in different jurisdictions.

This project aims to develop tools to help navigate this landscape and identify opportunities to deepen cooperation between governments and increase regulatory interoperability between their approaches.

The framework compares the approaches of five “early mover” governments: Canada, China, the EU, the UK, and the US, and contains seven categories: definition of AI, key aims, scope and focal areas, approach to risk, regulatory requirements (with deep dives into the role of technical standards, impact assessments, and audit), monitoring and enforcement, and flexibility and revisions.

This framework informs different stakeholders in their strategic planning:

Policymakers wanting to understand their regulatory options when designing a country’s approach or to contextualise their existing approach in relation to other jurisdictions;
International fora and organisations seeking to promote cooperation or convergence in governance between different jurisdictions;
Multinational corporations and SMEs trying to understand compliance requirements that may apply to them in different jurisdictions;
Prospective audit and certification bodies seeking to develop and provide bespoke AI auditing and certification services.

This report is a useful guide for anyone interested in forthcoming AI regulations and standards. Since AI doesn't carry the same conventional boundaries as other types of products and services, comparing key regions and their approach to AI oversight is important, as the use of these technologies is rapidly on the rise in all sectors and all regions of the world.

Cameron F. Kerry

Ann R. and Andrew H. Tisch Distinguished Visiting Fellow, The Brookings Institution, Centre for Technology and Innovation

AI research and development is extraordinarily collaborative across disciplines, institutions, and national borders. That makes collaboration among international partners essential to successful development of trustworthy AI, and understanding differences and commonalities in approaches to AI policy and development is an essential building block for this needed collaboration. This initial report from CEIMIA provides a framework for understanding that can be expanded to additional countries.

Ashley Casovan

Executive Director, Responsible AI Institute

Resources

Team

Stephanie King

CEIMIA

AI Initiatives Director

Collaborators

Cameron F. Kerry

Andrea Renda

Joshua P. Meltzer

Marjorie Buchser

Ashley Casovan

Rose Woolhouse

Surdas Mohit

Edward Teather

Marc-Etienne Ouimette

Alexandra Belias

Charlotte Stix

Yi Zeng

For this project, we are looking to expand our framework to other countries. We invite researchers and experts from countries with interesting regulatory development to join the project as advisors.

We also welcome funding opportunities to support the next phase of the project.

Contact us