info@businessghana.com
The inclusion of AI provisions in preferential trade agreements (PTAs) has been steadily rising. Since 2019, when the China-Mauritius Free Trade Agreement first mentioned AI, PTAs have progressively evolved to include elements addressing the broader implications of emerging technologies. While dedicated AI articles remain rare, AI-related provisions are often embedded in digital trade and data governance frameworks. These emphasize cross-border collaboration, ethical use of AI, algorithmic transparency and fostering trust in AI systems, demonstrating the potential of PTAs to drive not just economic growth, but also responsible and equitable technology adoption.
A global leader in PTAs with AI provisions
The recognition of AI in PTAs is growing in Asia and the Pacific. As of January 2025, 14 out of 16 trade agreements globally that incorporate AI provisions originate from economies within the Asia-Pacific region. Major contributors in the region include tech-advanced Asia-Pacific economies, such as Singapore, the Republic of Korea, Australia and New Zealand. These countries have enacted specific AI policies and relevant legal frameworks, such as data privacy, data flows, cybersecurity and intellectual property laws which impacts the development of AI. For example, in 2024, the Republic of Korea passed the AI Basic Act, a comprehensive legal framework on AI to take effect in January 2026. Moreover, they have also signed MoUs on AI cooperation, building upon their bilateral agreements, including those between Republic of Korea–Singapore (2022), Australia–Singapore (2024), and Australia–Republic of Korea (2024).
As expected, East and North-East Asia and Singapore lead AI-related developments in PTAs, while Least Developed Countries, Landlocked Developing Countries, and Pacific Island Developing Economies are generally absent (Figure 1). However, the Cambodia-UAE Comprehensive Economic Partnership Agreement stands out as a notable exception. While its AI-related provision remains at an early stage, it marks a significant first step for LDCs, which have traditionally lagged in adopting digital trade provisions.
Figure 1. Who are Active Players in AI-PTA Landscape, as of January 2025? (Source: ESCAP)
The current nature of AI provisions in PTAs
Owing to the cross-cutting nature of AI, the mentions of AI span various chapters, often within digital trade, economic cooperation, or innovation frameworks. Of the 14 Asia-Pacific PTAs, nine include provisions specifically focused on AI, while the remaining five incorporate AI within broader cooperation frameworks (Figure 2).
Figure 2. Where Is “Artificial Intelligence” Mentioned in Trade Agreements? (Source: ESCAP)
Although just 2.5 per cent of trade agreements globally explicitly reference AI, there is a growing emphasis on dedicated provisions that address ethical governance frameworks. The three key features of the dedicated AI provisions are:
Recognize the increasing importance of emerging technologies and/or AI, offering significant social and economic benefits to natural persons and enterprises.
Promote the development of internationally aligned emerging technologies and/or AI Governance Frameworks for ethical, trusted and responsible use. However, six out of nine PTAs with dedicated AI provisions do not specify which recognized frameworks should be used. Among the few that do, namely the Australia-United Kingdom PTA, Singapore-United Kingdom Digital Economy Agreement (DEA), and New Zealand-United Kingdom PTA, references are made to either the OECD Principles on AI (2019) or the Global Partnership on AI (2020), initiated by Canada and France in 2018.
Focus on cooperation on emerging technologies and/or AI, such as research sharing, responsible business use, commercialization opportunities and R&D investment.
Despite these shared elements, varying levels of comprehensiveness have been observed under the dedicated AI provisions. For instance, the Singapore-United Kingdom DEA and New Zealand-United Kingdom PTA further emphasize risk management, technological interoperability and technological neutrality in the governance frameworks.
Implications and way forward
Without coordination, AI-related trade policy risks becoming fragmented, reducing interoperability, and limiting economic growth. While the WTO is the logical forum to address the gaps, its slow progress has led countries to see PTAs as more immediate solutions.
Agreements in the Asia-Pacific region set early precedents for AI governance, but AI-related PTAs remain limited in scope and enforceability. Few countries are engaged in AI-related PTAs. Inclusive membership is key to ensuring more balanced and equitable AI trade rules. However, small developing countries currently face significant challenges to participate. AI trade provisions involve complex regulatory frameworks, cross-border data rules, and evolving global standards. Many developing countries still struggle to develop the necessary expertise and institutions for digital trade policies that support AI-driven trade and AI-driven growth with trust.
These challenges highlight the need for targeted assistance, regulatory cooperation, and capacity-building initiatives. ESCAP provides tools to bridge these gaps. For example, the Regional Digital Trade Integration Index (RDTII) helps countries assess their digital trade regulatory environment and align domestic policies with global and regional standards. LEGAL TINA enables trade negotiators to search and compare provisions across over 500 trade agreement texts, which is essential for developing informed negotiation strategies. And ESCAP’s Framework Agreement on Facilitation of Cross-border Paperless Trade in Asia and the Pacific provides an intergovernmental platform for countries to cooperate on AI solutions for electronic exchange and recognition of trade-related data and documents. Together, these and other tools may strengthen a country’s ability to engage in PTA negotiations on emerging digital economy issues, comply with commitments, and leverage international cooperation for sustainable domestic reforms and responsible AI use.
