The rapid evolution of artificial intelligence (AI) is reshaping global economies and societies, yet the existing frameworks for international trade remain ill-equipped to handle its complexities. As AI technologies increasingly rely on a global supply chain, including US-designed semiconductors and Taiwanese-fabricated chips, the absence of coherent governance poses significant risks. This lack of regulatory clarity could hinder innovation and create barriers to trade among nations.
While the World Trade Organization (WTO) has a limited mandate over digital services, it seems incapable of addressing the intricate landscape of AI. The political divides between the United States, European Union, and China complicate the possibility of a global AI treaty. As multilateralism retreats, bilateral Free Trade Agreements (FTAs) may need to serve as the testing grounds for AI governance, allowing nations to harmonize regulations and create predictable conditions for AI trade.
The AI supply chain is characterized by deep interdependence. At its core are semiconductor chips, primarily designed by leading firms in the United States but fabricated in Asia, particularly Taiwan and South Korea. These chips are critical for powering the data center networks that underlie large-scale AI models, predominantly operated by American hyperscalers. The models themselves are trained on vast datasets governed by a variety of legal frameworks, which may differ significantly from one jurisdiction to another.
This regulatory dissonance creates a patchwork compliance landscape. An AI product may be lawful in one country but non-compliant in another. The WTO has reported that only a small percentage of Regional Trade Agreements and Digital Economy Agreements explicitly address AI-related provisions. Most agreements merely acknowledge AI’s potential for social and economic benefits without instituting enforceable rules.
Initial attempts to develop a multilateral framework for digital technologies have stalled. The WTO’s procedural rules allow a single member to block any progress, and the growing polarization around AI governance has led some to advocate for the creation of sovereign AIs. Various organizations, including the OECD and UNESCO, have proposed ethical AI principles, but these remain non-binding and lack enforcement mechanisms.
Amidst this fragmented governance landscape, FTAs could emerge as pivotal instruments for shaping global AI norms. Although early agreements, such as the China–Mauritius Free Trade Agreement, made only passing references to AI, more recent agreements are beginning to incorporate the technology in more substantial ways. The Australia–Singapore Digital Economy Agreement, for instance, was the first to include a dedicated article on AI, albeit in an ‘endeavor-based’ format that lacks binding commitments.
As of October 2025, only 2 percent of trade agreements worldwide contain explicit AI-related provisions. These generally express broader intentions rather than enforceable standards. The language often intersects with other digital policy areas like cross-border data flows and data localization but stops short of addressing critical issues such as algorithmic accountability.
The need for comprehensive and enforceable provisions is evident. Future trade agreements should define key AI-related terminology to prevent uneven implementation. Terms such as “high-risk AI system” and “training data” should be explicitly clarified, with regular reviews to adapt to technological advancements. Moreover, agreements should transition from ‘endeavor-based’ language to include binding commitments, possibly through the establishment of a Committee on AI and Digital Trust within each agreement.
FTAs could also facilitate the harmonization of cross-border data governance by enabling trusted data flows while allowing countries to protect sensitive datasets. Predictable access to computing resources is essential for AI model development, and agreements should require parties to notify each other ahead of any export controls on critical AI-related infrastructure. Additionally, joint funding initiatives for public-interest AI projects could foster cooperation and innovation.
As countries navigate the complexities of AI governance, FTAs have the potential to serve as foundational frameworks for future multilateral cooperation. By experimenting with enforceable AI norms, nations can gradually build the trust necessary for broader agreements. The global AI landscape is evolving rapidly, and how nations choose to govern this transformative technology will have lasting implications for innovation, trade, and international relations.
The views expressed in this article belong to the authors and do not reflect the views of their employers or affiliated entities.
See also
Italy Orders Meta to Halt WhatsApp Terms Blocking Rival AI Chatbots Amid Antitrust Probe
Global Leaders and CEOs Unite at India AI Impact Summit 2026 to Drive Sustainable Innovation
Missouri Lawmakers Propose 11 Bills to Regulate AI Amid Growing National Concerns
Mistral AI Expands to Switzerland, Driving DeepTech Innovation and Funding Opportunities
AI Transforms Strategic Decision-Making: Enhancing Real-Time Insights and Reducing Bias in Business
