Technology Governance

Center for AI and Digital Policy, Michael Dukakis Institute for Leadership and Innovation
Boston, MA, Washington, DC
December 2020 :: 382 pages

Purpose and Scope
Artificial Intelligence and Democratic Values: The AI Social Contract Index is the first global survey to assess progress toward trustworthy AI. The AI Index 2020 has these objectives: (
1) to document the AI policies and practices of influential countries, based on publicly available sources,
(2) to establish a methodology for the evaluation of AI policies and practices, based on global norms, (3) to assess AI policies and practices based on this methodology and to provide a basis for comparative evaluation,
(4) to provide the basis for future evaluations, and
(5) to ultimately encourage all countries to make real the promise of AI that is trustworthy, human-centric, and provides broad social benefit to all.

Artificial Intelligence and Democratic Values focuses on human rights, rule of law, and democratic governance metrics. Endorsement and implementation of the OECD/G20 AI Principles is among the primary metrics. Opportunities for the public to participate in the formation of national AI policy, as well as the creation of independent agencies to address AI challenges, is also among the metrics. Patents, publications, investment, and employment impacts are important metrics for the AI economy, but they are not considered here.

The first edition of Artificial Intelligence and Democratic Values examined AI policies and practices in the Top 25 countries by GDP and other high impact countries. These countries are Australia, Belgium, Brazil, Canada, China, France, Germany, India, Indonesia, Italy, Japan, Korea, Mexico, Netherlands, Poland, Russia, Saudi Arabia, Spain, Switzerland, Taiwan, Thailand, Sweden, Turkey, United Kingdom, United States. High impact countries include Estonia, Israel, Kazakhstan, Rwanda, and Singapore.

Artificial Intelligence and Democratic Values will be published on an annual basis and will evolve as country practices change and new issues emerge.

:: The OECD/G20 AI Principles have Framed the Global Debate over AI policy. There are hundreds of frameworks for ethical AI, but only the OECD/G20 Principles have significantly shaped the policies and practices of national governments. Over 50 governments have formally endorsed the OECD/G20 AI Principles.

:: Governments have Both National Ambitions and Collaborative Goals. National AI policies typically reflect ambitions to be a leader in AI, to establish centers of AI excellence, and to promote economic growth. Many of these ambitions will set countries in competition for investment, personnel, and deployment. At the same time, countries recognize the need for global cooperation in such areas as public health, climate change, and sustainable development.

:: AI Safeguards Build on Data Protection Law. AI policy safeguards follows from other laws and policy frameworks, most notably data protection. The GDPR (Article 22), the Modernized Council of Europe Privacy Convention (Article 9), and the recently adopted California Privacy Rights Act in the US include explicit provisions for AI. The Global Privacy Assembly, the international conference of data protection officials, has recently adopted a sweeping resolution on the need for AI accountability.

:: Facial Surveillance as an AI “Red Line.” Few AI applications are more controversial than the use of AI for surveillance in public spaces. The use of facial recognition on a general population has raised widespread controversy with many NGOs stating it should be prohibited. Other controversial AI applications include the scoring of citizens, criminal sentencing, administrative service decisions, and hiring assessments.

:: Concern About Autonomous Weapons Remains. The risk of lethal autonomous weapons was among the first AI issues to focus the attention of government policymakers. Although many other AI policy issues have emerged in the last few years, concerns about autonomous weapons remains.

:: NGOs are Powerful Advocates for the Public. In Europe, civil society groups have published substantial reports on AI policy, documented abuses, and called for reform. Their advocacy has also strengthened democratic institutions which must now consider legal measures to address public concerns.

:: AI Policy is in the Early Days, but the Pace is Accelerating. AI research can be traced back to the 1950s but the effort of national governments to develop formal frameworks for AI policy is a recent phenomenon. Governments around the world are moving rapidly to understand the implications of the deployment of AI as more systems are deployed. We anticipate that the rate of AI policymaking will accelerate in the next few years.

1. Countries must establish national policies for AI that implement democratic values
2. Countries must ensure public participation in AI policymaking and also create robust mechanisms for independent oversight of AI systems
3. Countries must guarantee fairness, accountability, and transparency in all AI systems
4. Countries must commit to these principles in the development, procurement, and implementation of AI systems for public services
5. Countries must halt the use of facial recognition for mass surveillance