Singapore and France Show Two Roads to Governing Artificial Intelligence

Artificial intelligence has become the defining regulatory battleground of the decade, and a new comparative study published in the journal Global Public Policy and Governance offers one of the most detailed looks yet at how two very different states are wrestling with the same technological force. The research, conducted by Haris Alibašić of the University of West Florida and published on 4 June 2025, dissects the AI governance models of Singapore and France, two countries that have become emblematic of opposing philosophies: one built around innovation and voluntary trust-building, the other anchored in ethics, rights, and binding European law. By analyzing policy documents, regulatory frameworks, and implementation reports through systematic qualitative content analysis, the study reveals how national political culture, economic strategy, and the gravitational pull of international organizations shape the way governments translate abstract AI principles into enforceable practice.

The theoretical backbone of the analysis comes from Michael Zürn’s theory of global governance, which frames governance outcomes through three interlocking concepts: authority, legitimacy, and contestation. Alibašić supplements this with policy diffusion theory, which explains how regulatory ideas travel across borders, and multi-level governance theory, which captures how authority is distributed among local, national, and supranational institutions. This combination matters because AI governance is not written in a vacuum. Singapore’s framework responds to global standards set by bodies such as the Organisation for Economic Co-operation and Development, while France’s approach is inseparable from the European Union’s regulatory machinery. The study asks a deceptively simple question: when two states face the same technology and the same international recommendations, why do they produce such different governance architectures?

Singapore’s answer to AI governance is deliberately lightweight in legal form but heavy in practical infrastructure. The city-state’s National AI Strategy, launched in 2019 under its Smart Nation initiative, positions artificial intelligence as an economic engine that must be trusted before it can be scaled. Rather than imposing statutory obligations, Singapore’s Personal Data Protection Commission published a Model AI Governance Framework, now in its second edition, that offers companies voluntary guidance on transparency, explainability, and human oversight. The framework is paired with AI Verify, an open-source testing toolkit and governance system developed with industry through the AI Verify Foundation, which allows organizations to empirically demonstrate that their AI systems behave as claimed. This is governance by demonstration rather than by decree, and it reflects Singapore’s broader regulatory style of using sandboxes and voluntary codes to attract investment while preserving state credibility.

The Monetary Authority of Singapore extends this philosophy into the financial sector, one of the country’s most economically significant domains. Its Sandbox Express program gives firms a fast-track environment to trial AI-driven financial products under supervision, and its National AI Programme in Finance channels coordinated research into credit risk assessment, fraud detection, and customer service applications. The underlying logic is that trust in AI is built through verified performance, not through preemptive legal restriction. International observers, including the World Economic Forum, have highlighted Singapore’s approach as a model for balancing innovation with accountability, and the OECD’s AI Policy Observatory has documented how the model framework attempts to reconcile commercial dynamism with public confidence. In Zürn’s terms, Singapore seeks authority through technical competence and legitimacy through demonstrated outcomes, minimizing the contestation that hard law might provoke.

France, by contrast, has embedded AI governance within a rights-based constitutional and ethical tradition. The starting point is the 2018 Villani report, ‘For a Meaningful Artificial Intelligence,’ which argued that France and Europe needed a distinctive strategy grounded in human dignity, transparency, and public research capacity. That report catalyzed national investment and set the ethical tone for everything that followed. French governance then layered multiple advisory bodies into the process: the National Digital Council has assessed France’s competitiveness in AI, the National Pilot Committee for Digital Ethics issued a detailed opinion on generative AI systems, and the National Consultative Commission on Human Rights scrutinized facial recognition in public spaces. Each of these institutions feeds ethical deliberation into policymaking, creating a governance pipeline in which normative review precedes regulatory action.

The decisive external force on France’s model is the European Union. The European Commission’s 2019 Ethics Guidelines for Trustworthy AI established a soft-law baseline, and the 2021 proposal for a regulation laying down harmonized rules on artificial intelligence, which became the EU AI Act, converted those principles into binding, risk-tiered law. France, as an EU member state, must implement this framework, meaning its national AI policy operates within a multi-level governance structure where Brussels holds substantial regulatory authority. Scholars have described the EU’s capacity to export its standards globally as the ‘Brussels effect,’ and the study notes that this dynamic shapes not only France but also the incentives of non-EU states like Singapore, which must decide how much of the European approach to absorb in order to keep markets open and interoperability intact.

International organizations form the connective tissue between these two national projects. The OECD’s 2019 Recommendation on Artificial Intelligence established intergovernmental principles on human-centered values, transparency, and robustness that both Singapore and France reference in their policy documents. UNESCO’s 2021 Recommendation on the Ethics of Artificial Intelligence, adopted by 193 member states, pushed ethical AI further into the global mainstream, while the Global Partnership on AI and the World Economic Forum’s AI Governance Alliance created venues where governments, firms, and researchers negotiate shared norms. The study’s content analysis shows that these organizations function as norm entrepreneurs in the sense described by Finnemore and Sikkink: they articulate standards, diffuse them through reports and summits, and confer legitimacy on states that comply. Yet the diffusion is not mechanical. Each country filters international norms through domestic political and economic priorities, producing hybrid governance rather than uniform convergence.

This is where the comparative findings become most striking. Singapore’s innovation-driven model and France’s ethics-centric model are not simply opposites; they represent different resolutions of the same legitimacy problem. Singapore resolves it by proving trustworthiness through testing regimes and voluntary frameworks, betting that flexible governance will outpace rigid rules in a fast-moving field. France resolves it by grounding AI in fundamental rights and democratic deliberation, betting that legitimacy derived from ethical and legal authority will produce more durable public acceptance. The study finds that both approaches face contestation. Singapore’s voluntary model invites criticism that it lacks enforceable teeth when systems cause harm, while France’s rights-based model faces pressure from industry over compliance costs and from the pace of EU rulemaking. Neither country has escaped the fundamental tension between fostering innovation and constraining risk.

The paper’s central conclusion is that effective global AI governance demands a polycentric approach, drawing on Elinor Ostrom’s insight that complex collective-action problems are often best managed by multiple, overlapping centers of authority rather than a single hierarchical regulator. In practice, this means universal ethical principles, such as those articulated by the OECD and UNESCO, must be implemented through localized strategies that respect each jurisdiction’s legal traditions, economic structure, and cultural expectations. A polycentric system also allows experimentation: Singapore’s testing toolkits and France’s ethics committees generate different kinds of evidence and legitimacy, and the global system benefits when these innovations diffuse across borders. The alternative, a single monolithic global AI regulator, appears both politically unattainable and practically ill-suited to a technology whose risks and benefits vary enormously across contexts.

For policymakers, the study’s implications are immediate. As the EU AI Act enters into force and other jurisdictions draft their own rules, the Singapore-France comparison offers a template for diagnosing trade-offs: voluntary frameworks trade enforceability for agility, while statutory regimes trade agility for accountability, and both depend on international standards for coherence. For scholars, the research demonstrates the analytical value of combining global governance theory with policy diffusion and multi-level governance perspectives to explain why international AI norms land differently across states. As artificial intelligence systems grow more capable and more embedded in finance, healthcare, and public administration, the question raised by this research will only intensify: how the world harmonizes AI governance without erasing the national diversity that makes governance legitimate in the first place.

Subject of Research: Comparative analysis of artificial intelligence governance policies in Singapore and France

Article Title: Harmonizing artificial intelligence (AI) governance: A comparative analysis of Singapore and France’s AI policies and the influence of international organizations

Article References: Alibašić, H. (2025). Harmonizing artificial intelligence (AI) governance: A comparative analysis of Singapore and France’s AI policies and the influence of international organizations. Global Public Policy and Governance, 5(2), 93-113. https://doi.org/10.1007/s43508-025-00116-w

Image Credits: AI Generated

DOI: 10.1007/s43508-025-00116-w

Keywords: artificial intelligence, AI governance, Singapore, France, global governance, European Union, policy diffusion, OECD, UNESCO, multi-level governance, AI ethics, regulation