Where Is Singapore’s AI Regulation Headed?
Is Singapore “pretty relaxed” when it comes to regulating artificial intelligence? Haslinda Amin—Bloomberg Television’s chief international correspondent for Southeast Asia—described it this way during a panel at the 2025 meeting of the World Economic Forum in Davos, Switzerland. In response, Gan Kim Yong, the deputy prime minister of Singapore, countered that the city-state has “an elaborate code of conduct” for how to develop and deploy AI in an ethical manner, while applying a “very light touch in terms of regulation” to allow for flexibility as the technology evolves. Gan summed up Singapore’s approach as “nimble” and “flexible—but at the same time, we are very careful in providing guidance and guardrails.”
Singapore does not have laws and regulations specifically addressing AI.
Singapore does not have laws and regulations specifically addressing AI, hence Amin’s characterization. The assumption here is that a legal framework such as the European Union’s Artificial Intelligence Act, passed in 2024, is the gold standard. Some countries, including Brazil and South Korea, have followed the European Union in adopting a single, overarching law focusing on restricting high-risk applications of AI; other nations, such as China, have opted for a range of focused rules—for example, requiring AI-generated content to be labeled. Meanwhile, Singapore, like many other nations, relies primarily on existing laws and regulations that address common consumer and societal harms and supplementing those with voluntary guidelines. However, given the risks that some AI applications pose, it is unclear whether this is sufficient.
Nevertheless, as a self-proclaimed “Smart Nation,” Singapore has worked hard to earn a position on the frontier of the global conversation around governing and regulating AI. To this end, it launched a voluntary Model AI Governance Framework (MAIGF) in 2019, which provides guidance to the private sector on ethical and governance issues related to the deployment of AI. This was followed by a National AI Strategy later that year. In May 2024, a companion document to the MAIGF focusing specifically on generative AI was published. Meanwhile, the National AI Strategy was updated in 2023, promoting a new vision of “AI for the Public Good, for Singapore and the World,” which focuses on building Singapore’s excellence in AI and empowering individuals and businesses to use AI.
Related Reading
Even more importantly, in 2022 the city-state went beyond studies and papers and got involved in the practicalities of implementing responsible AI by launching AI Verify. AI Verify is a governance testing framework and toolkit that applies 11 ethics principles—including transparency, explainability, safety, security, robustness, and fairness—to evaluate the performance of AI systems. It includes a software component to conduct technical tests for fairness, explainability, and robustness. AI Verify is a voluntary program intended to build public trust by providing criteria to evaluate how well AI systems align with ethical principles.
Singapore is home to more than 6 million people packed into about 734 square kilometers, but this “little red dot,” as it’s often referred to, has a long history of successfully cultivating bilateral and business relationships to safeguard its position as a commercial and technological hub. In AI, Singapore’s efforts appear to be paying off, and it has steadily advanced its credentials as a key international player in AI governance. In 2023, the country was one of 29 signing the Bletchley Declaration at the inaugural AI Safety Summit hosted by the United Kingdom. And at the most recent AI Action Summit held in Paris in February 2025, Singapore launched the Global AI Assurance Pilot, which aims to establish best practices related to technical testing of generative AI applications, along with several other initiatives.
Singapore’s position on AI governance takes into account the city-state’s considerable constraints around land, energy, and population, while positioning it to be a leader in developing AI systems. Between 2013 and 2024, Singapore ranked 10th among countries receiving private investment in AI, with $7.27 billion invested. By one measure of its ability to innovate in AI, Singapore ranked seventh overall in 2023, with 0.98 granted AI patents per 100,000 inhabitants.
All this points to a strong and hopeful future for Singapore when it comes to AI. But despite its proactive approach and leadership position, it remains to be seen whether Singapore’s approach can address the wide range of serious risks associated with AI, particularly those affecting safety and security. There are three distinct challenges ahead that lack straightforward solutions, and how Singapore addresses them will be important not only for its citizens, but for the rest of the world as well.
Singapore’s position on AI governance takes into account the city-state’s considerable constraints around land, energy, and population, while positioning it to be a leader in developing AI systems.
The first and perhaps most daunting of these challenges will be to safely navigate between China and the United States. Both countries view dominance in AI as a determinant of global leadership, and each will leverage AI in its effort to prevail over the other. Singapore has traditionally walked a tightrope, cultivating strong ties with both superpowers. But today, each step forward seems more perilous than the last. If recent history is a guide, we can expect continued uncertainty and inconsistency as the Trump administration’s policy agenda unfolds. China, navigating economic headwinds amid a rapidly aging population, faces uncertainty, too. However, China sees itself evolving into an AI “supermarket” driven by its AI Plus Initiative, which aims to integrate AI into traditional sectors of the economy, such as agriculture and manufacturing.
It will be a challenge for Singapore to concurrently maintain openness to the competing technology ecosystems of China and the United States as they increasingly wall themselves off from one another. There has been a consistent refrain, particularly in Southeast Asia, regarding not being forced to choose between the superpowers, but it is impossible to ignore China’s economic interdependence with the region. China is the most significant trading and investment partner for many Southeast Asian countries, and it has been a reliable source of affordable technology that supports development and economic growth. In January 2025, Indonesia joined BRICS, an intergovernmental organization founded in 2009 between Brazil, Russia, India, and China with the aim of countering Western dominance. BRICS now has ten member states, and Malaysia, Thailand, and Vietnam have also indicated interest in joining. This suggests that the regional order that has long centered on the Association of Southeast Asian Nations, or ASEAN, which was founded during the Cold War to resist communism, could be unravelling.
The second challenge will be to navigate the domestic politics surrounding Singapore’s identity as a Smart Nation, which will shape its approach to AI governance on regional and global platforms. Despite Singapore’s desire to be a Smart Nation, the city-state’s leadership is well aware that digitalization comes with social costs. Of particular concern is the inequality and exclusion fostered by the creation of a digital divide, which has left many elderly Singaporeans ambivalent. Furthermore, as Singaporeans’ daily lived experience is increasingly mediated by technology, the potential for cybersecurity threats, scams, and manipulation of the information environment also intensifies. AI has been touted as a solution to some of these challenges, but in a country that has long prized social cohesion, it could also present a danger to society by polarizing and alienating people from each other.
Thus, a particular focus around AI regulation is that it should protect individuals from harm. For example, many people are concerned about job displacement in countries with high labor costs, like Singapore, as companies adopt AI to maximize profitability. In February 2025, Singapore’s largest bank, DBS, announced that it would cut some 4,000 temporary and contract roles in the coming three years as it integrates AI into its operations. The bank also expects to create 1,000 new AI-related jobs, but it is unclear to what extent those who are displaced can adapt to these roles.
Too little attention has been devoted to understanding and mitigating potential harm from the intentional misuse or manipulation of AI.
Singapore’s third challenge will be to ensure the safety and reliability of AI systems as they are adopted. Much of the conversation about governance of AI focuses on unintentional harm rising from bias in the design and training of AI models. However, too little attention has been devoted to understanding and mitigating potential harm from the intentional misuse or manipulation of AI.
Adversarial attacks on AI systems pose a significant threat, as they are difficult to detect and measures to mitigate them have yet to be meaningfully realized. Meanwhile, increasing adoption of AI generates vulnerability to these attacks. In a scenario where AI underpins the safe and reliable operation of systems such as autonomous vehicles or access control systems relying on face recognition, attacks to fool these systems could have dire economic, social, and safety consequences.
A variety of institutions are conducting important work to address adversarial attacks on AI. One useful starting point is a taxonomy released by the US National Institute of Standards and Technology in January 2024 of the various ways in which AI systems can be attacked and possible ways to mitigate these attacks. And some existing cybersecurity governance mechanisms can be extended to AI systems to reduce the time and resources needed to implement mitigation measures as they are developed.
In the years to come, Singapore will have to navigate many other challenges with AI, and it seems unlikely that the city-state will be able to remain “pretty relaxed” about regulation. As approaches to AI governance evolve in other countries and regions, Singapore must pay close attention and continue to refine its own policy positions and strategies. A particular difficulty will arise when developments in the world of AI cause turbulence beyond the reach of existing governance levers, such as the potential rise of artificial superintelligence. Some forecasts predict this scenario to be only a few years away. Nevertheless, although countries have a range of assessments and opinions regarding the probability of such forecasts, they will need to engage in some form of preparation in the present. Singapore has assiduously avoided getting caught up in the debate around the rise of artificial superintelligence thus far, preferring instead to focus on what it can pragmatically address at the current state of AI’s technical advancement. It remains to be seen whether this approach will be sufficient.