Today, after extensive conversations with governments, academia and civil society, we’re releasing 10 AI Policy Gold Standards that offer a practical roadmap to help emerging economies build AI-ready infrastructure, prepare people for the AI opportunity, and govern AI responsibly.
According to Goldman Sachs, AI could increase global GDP by 7% over a ten-year period, but that’s dependent on widespread adoption. AI is already creating an unprecedented opportunity to accelerate growth and development for all. That’s why in 2024 Google launched the AI Sprinters Framework, which outlined strategies for governments to pursue AI transformation.
Policymakers increasingly recognize that governments that seize the AI opportunity will both define digital leadership and accelerate their own development. The most common request across our conversations was to build on the AI Sprinters Framework with an actionable roadmap, including policy recommendations designed to help capture the full potential of AI. To do that, we’ve synthesized our recommendations into 10 AI Policy Gold Standards, broken down into the three main phases in a nation’s AI transformation: building an AI-ready ecosystem, achieving broad-based AI adoption, and creating an enabling policy environment.
Innovation infrastructure — build the future-ready AI ecosystem
1. Enhance cloud capacity and implement cloud-first policies. Cloud computing is the gateway to harnessing AI’s power. Governments should prioritize enhancing national cloud capacity and adopting “Cloud-First” policies. For example, Singapore has migrated over 80% of eligible government workloads to the public cloud.
2. Make public sector data available. Access to high-quality, open datasets is vital for AI development. Governments should establish centralized, open-sourced national data repositories — as with the INDIAai mission of creating a centralized repository of AI datasets across critical sectors.
3. Facilitate access to datasets in sensitive contexts. AI-enabled research in areas like healthcare can save lives. Governments should facilitate safe access to sensitive datasets. The Provincial Health Data Centre (PHDC) in South Africa consolidates person-level health data to provide approved researchers with anonymized datasets, ensuring privacy by design.
4. Invest in a robust AI ecosystem. To ignite a vibrant AI ecosystem, governments should implement a startup-friendly policy environment, encourage public-private partnerships, and attract investment in start-ups and AI transformation. Rwanda’s National AI Policy facilitates collaboration with the private sector to co-invest in local AI companies.
Broad-based adoption — drive national AI diffusion across all sectors
5. Facilitate government adoption of AI. Governments should lead by integrating AI into their operations to improve public service efficiency. The state of Minas Gerais, Brazil, uses Google Cloud AI to streamline environmental violation report processing, analyzing data and legal precedents to enhance efficiency.
6. Facilitate SMBs’ use of AI. Small and medium-sized businesses (SMBs) are the largest employers in most economies, and AI is a powerful tool to boost their productivity. Governments should help entrepreneurs access and adopt AI solutions through grants and skilling programs. The Australian government’s AI Adopt Centres help SMBs adopt AI services responsibly and grow their businesses.
7. Implement a comprehensive AI Skills Plan. Ensuring all citizens are ready for the AI era is crucial and requires an approach that serves all — from workers to civil servants, and developers. The UAE’s National AI Strategy provides training to the public, government employees and STEM students. The UAE Prime Minister’s Office, in partnership with Apolitical, launched AI courses for 54,000 public servants. This is part of a Google.org grant to train one million government employees globally.
Enabling policy environment — ensure long-term success
8. Adopt international AI standards in domestic regulation. To avoid regulatory fragmentation and promote global interoperability, governments should adopt international AI standards. Risk management standards like ISO 42001 provide a strong technical foundation for responsible AI development and deployment.
9. Create or maintain copyright and privacy frameworks that enable the use of publicly available information. A balanced copyright framework is essential to support the responsible development of AI — opening access to the data needed to train these systems — while still ensuring that rightsholders can protect their creative works. Governments should support copyright laws that enable training on publicly available content through relevant limitations and exceptions like text and data mining exceptions, as seen in Singapore, Japan and the EU.
10. Leverage existing regulations as a starting point. To avoid a fragmented regulatory landscape, governments should first assess how existing regulations already apply to AI applications before introducing new, AI-specific regulations. Israel’s AI Program suggests using “soft” regulatory tools and empowering sectorial regulators to address AI risks, to allow for incremental development of the regulatory framework.
A roadmap for AI Leadership
AI adoption is a national imperative — and our hope is that these 10 AI Policy Gold Standards offer a practical blueprint for realizing its potential. We’re committed to partnering with governments, academia and civil society to fuel innovation and develop safe, responsible AI solutions that work for all.
