Key Findings from the Artificial Intelligence and Democracy Values Index

April Yoder is Editor-in-Chief of the Artificial Intelligence and Democratic Values Index 2025, and Christabel Randolph is Associate Director at the Center for AI and Digital Policy (CAIDP).

With the recent thrust by the Trump Administration to ensure US global AI leadership, the questions surrounding DOGE’s use of AI to analyze agency operations for waste and fraud, and the ongoing struggles over the European Commission’s General Purpose AI Code of Practice reveal the challenges of implementing and enforcing policies designed to ensure AI systems serve the public interest, are human-centered, and aligned with democratic institutions and the rule of law. There is now a near-global consensus and convergence on governance pillars like transparency, accountability, and fairness for AI systems. Yet the pace of implementation and enforcement of these safeguards remain the key challenges to address in the future.

The Center for AI and Digital Policy (CAIDP) annual report, Artificial Intelligence and Democracy Values Index (AIDV), offers a measure of how far we have come in AI policy and governance and some of the immediate and longer-term challenges that lie ahead. With reports on the AI policies and practices for 80 countries on six continents, the AIDV truly reflects the global landscape. CAIDP’s team of AI policy researchers represent 120 countries; the 2024 cohort represented more than 80. This global coverage and focus on fundamental rights, democratic values, and the rule of law make the AIDV 2025 unique among AI indexes. The AIDV provides breadth in covering current policies and practices as well as a historical record of developments.

It would be easy to assume that AI governance is more mature and vibrant in the EU and North America. Significantly, AIDV shows that 64% of the countries with meaningful AI policies are within Africa, Asia, Latin America & Caribbean, Middle East, and Oceania. The tiered rankings within the AIDV also reveal that countries within these regions have progressed more in ensuring ethical AI policies while Canada, Japan, Korea, Netherlands, and United Kingdom lead the rankings.

The newest edition of the report, AIDV 2025, revealed 2024 as a watershed year for AI policy, marked by the culmination of years of negotiations in global and regional bodies in agreements such as the Council of Europe AI Treaty, African Union Continental AI Strategy, and EU AI Act. Here, we review some of the key developments of 2024 and challenges and next steps to ensure progress in 2025.

Despite challenges in implementation and enforcement, CAIDP notes near-global consensus on the principles for AI development. The first ever international treaty on AI, the Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and Rule of Law went into force and is open for signature by member and non-member states. So far, 41 countries are signatories to the AI Treaty, including the United States, the United Kingdom, EU member countries, Japan, and Canada.

The widespread support for UN Resolution L.49 on Seizing the Opportunities of Safe, Secure and Trustworthy Artificial Intelligence Systems for Sustainable Development, with more than 120 countries co-sponsoring reflected this consensus for AI systems that are “human-centric, reliable, explainable, ethical, inclusive, in full respect, promotion and protection of human rights and international law, privacy preserving, sustainable development oriented, and responsible” (p. 2). These principles are established in global frameworks such as the 2018 Universal Guidelines for Artificial Intelligence, OECD AI Principles (2019, 2024 update), and UNESCO Recommendation on the Ethics of Artificial Intelligence, which figure prominently in CAIDP’s evaluation of the AI policies and practices in the 80 countries included in the report.

Member countries took steps to implement these principles and the affirmation in the General Assembly that “The same rights that people have offline must also be protected online, including throughout the life cycle of artificial intelligence systems” at the Summit of the Future in September 2024, which produced the Pact for the Future and Global Digital Compact. The Global Digital Compact centers on five objectives to ensure AI and other technological developments benefit everyone. Countries and organizations that endorse the compact agree to 1) Close digital divides and accelerate progress toward the SDGs; 2) Expand inclusion in and the benefits from the digital economy; 3) Develop open, safe, and secure digital spaces to respect, protect, and promote human rights; 4) Foster trustworthy data transfer through responsible governance; 5) Govern AI for the benefit of humanity.

The African Union Continental Artificial Intelligence Strategy, adopted in July 2024, focuses on harnessing AI for people while addressing risks to people and the environment. The Strategy centers sustainable development goals that build on existing capacity as well as investments in human development through education and skills-building. The Strategy demonstrates African countries’ commitment and that of the African Union to becoming participants in AI governance and development rather than consumers or suppliers of natural resources. The infrastructure investments outlined in the Strategy reveal the reach of AI governance beyond technical and economic development to democratic, cultural, and social institutions.

The EU AI Act was approved in early 2024, and the first provisions entered into force in August 2024. EU member states and partners in the European Economic Area are working to align national regulations with the Act and to name authorities to oversee and enforce the Act. With enforcement of the Act set to begin in August 2025, CAIDP and other organizations have been active to ensure the Codes of Practice uphold the principles and protections in the Act. Already, civil society organizations are sounding the alarm on the Code of Practice consultative process and the risk of watering down the protections set out in the EU AI Act.

The AIDV shows mixed results for AI legislation and policies in other countries. Global majority countries in Latin America, MENA, Asia, and Africa progressed most notably in implementing the UNESCO Recommendation on AI Ethics by completing the Readiness Assessment Methodology (RAM). The RAM assesses a country’s readiness along legal, socio-cultural, economic, technical, and infrastructural dimensions to ethically adopt AI and implement safeguards. Since the release of the AIDV, the Netherlands and Cuba have joined Brazil, Chile, the Dominican Republic, Mexico, and Uruguay from Latin America; Indonesia from Asia; Morocco and Saudi Arabia from MENA; and Gabon, Kenya, Mozambique, Senegal, and South Africa from Africa in completing the RAM.

In 2024, Canada’s Artificial Intelligence and Data Act (AIDA) and the Digital Charger Implementation Act of which it was a part failed to pass Parliament. The Korean AI Framework Act, passed in January 2025, offers another example of AI-centered legislation. The Korean legislation establishes obligations for the use of AI systems in high-impact sectors such as healthcare, but falls short of prohibiting AI uses as in the EU AI Act. AIDV 2025 offers more detail on national developments, such as policies in Costa Rica and findings of the RAM.

Beyond national or regional legislation, international bodies such as the G7 and Association of South East Asian Nations (ASEAN) have developed tools to assist nations, organizations, and corporations in developing AI according to shared principles. The first reports from the G7 Hiroshima AI Process (HAIP) Reporting Framework, based on the HAIP International Code of Conduct for AI developers, are now available for review through the dedicated OECD AI Policy Observatory page. The ASEAN Guide on AI Governance and Ethics offers organizations a tool for meaningful, human-centric development aligned to local needs and values.

CAIDP’s efforts in 2025 seek to consolidate the gains of 2024 and add to them. Our advocacy and advice to international organizations and countries will center on global adoption of the international AI Treaty; ensuring all people have legal protection to algorithmic transparency, including access to contest adverse outcomes; prohibitions on systems that violate human rights and democracy values; human oversight across the AI lifecycle, including a termination obligation; and establishing clear liability rules for AI systems and the entities who develop and deploy them. The AIDV is a global review as well as a global effort toward awareness, capacity, and inclusion for AI policies that center people and the public interest. It is a resource and call to action to policymakers that proclamations and principles must be reflected in practices to ensure that AI does not contribute to the erosion of the fundamental safeguards for people’s rights and well-being.