International Trade and Artificial Intelligence: Is trade policy ready for Chat GPT?

This research paper by Pascal Krummenacher analyzes the gaps in current international trade policy caused by the rise of artificial intelligence (AI). It argues that existing World Trade Organization (WTO) frameworks—specifically regarding the goods-services distinction, professional certifications, services delivery modes, and intellectual property—are ill-equipped for AI. While acknowledging the WTO's slow pace and the lack of global normative consensus, the author suggests that the organization should utilize 'soft law' to coordinate international AI regulation and prevent fragmentation.

• AI creates significant regulatory gaps in the distinction between goods and services. As AI is integrated into physical products like robotics and self-driving cars, it becomes unclear whether the General Agreement on Tariffs and Trade (GATT) or the General Agreement on Trade in Services (GATS) should apply. The current reliance on a 'patchwork' of free trade agreements to resolve this increases trade costs.
• Existing GATS frameworks are ill-suited for AI in two ways: first, professional certification and legal personhood requirements may exclude AI systems (such as Chat GPT or Harvey) from market access commitments; second, the four existing modes of delivery fail to account for services embedded in goods. This leads to 'tariffication' of services, where software installed on hardware is taxed via customs value, whereas the same software delivered online is not. This has led to proposals for a 'mode 5' of services delivery, which a 2017 study suggested could increase world trade by EUR 500 billion.
• The WTO's Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS) lacks a definition for AI-generated works. This has resulted in inconsistent domestic laws among member states, with some providing full protection and others requiring human creativity, creating an unsatisfactory environment as AI-generated intellectual property trade grows.
• While the WTO is the logical venue for addressing these gaps to avoid international fragmentation, it faces a 'pacing problem' where technology evolves faster than regulation. Additionally, divergent national interests—such as the different privacy approaches of the US and EU, and the differing budget needs of rich versus poor countries—hinder hard law agreements. Consequently, the author recommends the use of 'soft law' (informal rules and norms) as a flexible mechanism to coordinate AI regulation.

APA

Krummenacher, P. (2023). International Trade and Artificial Intelligence: Is trade policy ready for Chat GPT? International Institute for Sustainable Development. https://www.iisd.org/articles/policy-analysis/international-trade-artificial-intelligence-chatgpt

Chicago

Krummenacher, Pascal. International Trade and Artificial Intelligence: Is trade policy ready for Chat GPT? International Institute for Sustainable Development, 2023. https://www.iisd.org/articles/policy-analysis/international-trade-artificial-intelligence-chatgpt.

Wikipedia

{{cite report |last1=Krummenacher |first1=Pascal |title=International Trade and Artificial Intelligence: Is trade policy ready for Chat GPT? |publisher=International Institute for Sustainable Development |date=14 April 2023 |url=https://www.iisd.org/articles/policy-analysis/international-trade-artificial-intelligence-chatgpt |access-date=17 August 2026 |via=Climate Insights Directory}}

BibTeX

@techreport{krummenacher2023international, author = {Krummenacher, Pascal}, title = {{International Trade and Artificial Intelligence: Is trade policy ready for Chat GPT?}}, institution = {International Institute for Sustainable Development}, year = {2023}, month = apr, url = {https://www.iisd.org/articles/policy-analysis/international-trade-artificial-intelligence-chatgpt}, urldate = {2026-08-17}, note = {Indexed by Climate Insights Directory} }