6
AI to boost global trade by 40 per cent; Singapore cited as regulatory model: WTO report

AI to boost global trade by 40 per cent; Singapore cited as regulatory model: WTO report

According to the report by the World Trade Organisation, Singapore is an example of how resource-constrained economies can effectively leverage AI by adapting existing open-weight models rather than developing proprietary ones from scratch.

Cargo ship transporting containers at sea with digital graphics overlay representing global trade and logistics technology.

The Republic has been highlighted as a frontrunner for its proactive approach to artificial intelligence (AI) regulation in a World Trade Organization (WTO) report, which projects that AI will boost global trade by nearly 40 per cent by 2040.

In the report, released on Wednesday (17 Sep), WTO estimated that AI could generate global gross domestic product increases of 12 to 13 per cent by 2040, with global trade to rise by 34 to 37 per cent across different scenarios.

The largest growth will be in the trade of digitally deliverable services, including AI services, which could see a 42 per cent rise. This increase stems from reduced operational costs, strong growth in highly tradable AI services concentrated in key regions, and above-average productivity gains in digitally deliverable services.

But the WTO warned that the benefits of AI are not guaranteed, and could deepen existing economic and technological divides without concerted international action.

“Without proactive policy responses and greater international cooperation, AI could deepen inequalities rather than reducing them,” said WTO director-general Ngozi Okonjo-Iweala.
 


Regulation

On regulatory frameworks, the report identified Singapore alongside the United Kingdom and the United States as adopting a principles-based approach to AI regulation. This means regulation relies on non-binding principles, often supported by technical standards and cross-sectoral regulations, rather than strict prescriptive laws.

“While recognising AI-related risks, they view strict regulation as premature given AI’s ongoing evolution,” said the report.

In contrast, jurisdictions following a rules-based approach, such as China, Brazil, and the European Union, are introducing specific AI legislation to ensure regulatory clarity, enable enforcement against unlawful AI deployment, and protect consumers from potential harms.

Singapore was also recognised for its practical approach to AI development and safety. The city-state has established an AI safety institute to build expertise and manage risks, placing it among advanced economies leading such initiatives.

More notably, Singapore was highlighted as an example of how resource-constrained economies can effectively leverage AI by adapting existing open-weight models rather than developing proprietary ones from scratch.

Open-weight models refer to AI systems with publicly accessible parameters that can be modified by developers. The WTO cited the Infocomm Media Development Authority’s GPT-Legal, an AI-powered legal research tool built on Meta’s Llama model.

“Fine-tuned to the Singapore legal context, including historical case law, GPT-Legal was developed using significantly fewer resources and in a shorter time frame than many large proprietary models,” said the WTO.

“This demonstrates the potential for developing economies to adapt existing models for local needs, offering a resource-efficient way to build context-specific AI applications.”

This approach contrasts with AI powerhouses such as the United States and China, which remain “clear leaders” in developing original models. As at April 2024, the United States had developed 439 generative AI models, followed by China with 117 and the United Kingdom with 88, while Germany, France and Canada also made “significant contributions”.
 

Asia’s role

Asian economies are playing an increasingly central role in global AI trade, with several East Asian and Southeast Asian countries recording rising imports of AI-related goods as they integrate deeper into production networks.

China and the United States lead as the largest importers of AI-enabling goods, while Singapore, Taiwan, South Korea, Japan, and Malaysia have also seen growing imports reflecting their expanding role in the AI supply chain.

On the export side, activity is concentrated in East Asia, with China, Taiwan, and South Korea leading particularly in intermediate inputs and equipment.

There are also opportunities in AI dataset training, a market tipped to grow to US$3.6 billion in 2025, and to US$17 billion by 2032, from US$2.9 billion in 2024.

North America currently dominates this market, but the Asia-Pacific region is projected to experience the highest growth rate between 2025 and 2032, said the report.
 


Cost and benefits

At a media conference, Johanna Hill, WTO deputy director-general, noted that uncertainty in trade policy weighs heavily on business confidence, investment, and supply chains.

But while the trade landscape is changing fast and geopolitics dominate the headlines, there are other trends that are reshaping the future of the global economy and international trade. “AI is certainly one of them,” said Hill.

The report found that AI could be a bright spot for trade, offering opportunities to reduce trade costs, boost productivity, and expand participation in global markets.

A survey conducted by the WTO and the International Chamber of Commerce found that nearly 90 per cent of companies using AI reported tangible benefits in trade-related activities, and more than 70 per cent anticipate trade cost savings.

While real wages are expected to rise, the skill premium – the ratio of high-skilled to low-skilled wages – is projected to decline moderately, by 3 to 4 per cent, as AI substitutes for medium and high-skilled tasks.

This means that there is a risk of job displacement, particularly for middle-skilled workers, said the WTO.
 

Source: The Business Times © SPH Media Limited. Permission required for reproduction.

Related Content

Subscribe Icon
The latest business insights and news delivered to your inbox
Subscribe now