Why AI is the new frontier global trade must learn to cross

Global trade is in constant flux, reshaped by technological breakthroughs that open new frontiers for commerce but also present fresh challenges for international trade policy. This was the case with the advent of the internet and the propagation of e-commerce, and it will be so with artificial intelligence. Given the sometimes-glacial pace at which multilateral policy-making moves, and the generalized and profound impact that AI will have on trade, there is an urgent need to build a policy architecture that reconciles trade and AI.

This is the subject of a recent white paper published by Penelope Naas and Usman Ahmed. In it, they explore how AI is reshaping trade, the innovations and efficiencies gained, as well as the policy challenges it presents for organizations such as the World Trade Organization (WTO).

There is great promise in the benefits that AI can bring. Some estimates suggest that it could raise global GDP by 7% over 10 years. However, innovation must be balanced against risks related to privacy, bias, cybersecurity and labour markets. As such, several policies have already been enacted at both an international and domestic level to navigate the costs and benefits of AI, many of which touch on trade.

At an international level, the OECD was an early mover. In 2019, it adopted the OECD AI Principles, the first intergovernmental standard on AI that was adopted by 47 countries. It provides value-based principles and policy recommendations for both government and other AI actors. UNESCO released the Recommendation on the Ethics of Artificial Intelligence in November 2021 to address the ethical risks associated with AI. In 2023, the UK government convened officials, scientists and other experts at an AI Safety Summit and released the Bletchley Declaration for AI to jumpstart international collaboration on the safety and regulation of “frontier” AI.

At a regional level, the European Union passed one of the most comprehensive legal frameworks on AI in March of this year. The EU AI Act takes a values-based approach, aiming to foster AI innovation, while upholding human rights, the rule of law and democracy. The act also contains provisions that ensure AI systems are in line with its broader trade commitments. Additionally, six bilateral trade agreements around the world already include provisions for AI.

Despite these efforts, there are several pressing concerns regarding AI’s impact on international trade.

1. Fragmentation between domestic policy: The various domestic regulations on AI emerging in many countries, including major trading actors such as the US, China, Australia, Singapore, Japan and others, risk causing regulatory fragmentation: different rules in different regions that hinder international trade. For example, an AI company that operates out of a country where data privacy laws are less stringent may face legal challenges if it begins to operate in a country with stronger data laws. The WTO’s Technical Barriers to Trade committee looks at obstacles that result from domestic regulation and may need to expand its scope to further address barriers caused by AI.

2. Implications for the General Agreement on Trade in Services (GATS): The GATS ensures that WTO members conduct fair and open exchange of services (financial services, insurance, travel etc.). GATS classifies service trade into four different “modes”: cross-border trade, consumption abroad, commercial presence and presence of natural persons. AI could qualify under several of these; the distinction is important because under the agreement, countries have varying obligations based on the mode. Therefore, mode ambiguity can cause confusion and make it difficult for countries to fulfil trade commitments.

3. Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS): The TRIPS agreement sets standards for copyright protections on intellectual creations. It stipulates that computer programs should be protected in the same way as literary works. However, under the Berne Convention, only humans can be considered authors. This leads to questions around who bears the rights to AI-generated materials and who can trade in them. TRIPS needs updating to clarify if AI-generated works are copyrightable, and how this protection might differ from traditional copyright.

4. Subsidies: WTO subsidy rules apply to government support for producers. While current rules cover goods used in AI production, such as semiconductors, they don't fully cover the actual AI service. Therefore, the scope to regulate subsidies on AI software, source code and algorithms is limited. This challenges the tenant of fair competition in trade, as large countries can pour resources into AI development, while the rest of the world may struggle to keep up.

Despite current impasses at the WTO on plurilateral and multilateral negotiations, there are a number of steps that the organization can take to address AI’s impact on trade:

1. Committees: The WTO could use existing committees to explore issues pertaining to AI. The aforementioned Technical Barriers to Trade Committee has already been used by member states as a forum to discuss the technology. By using this committee or establishing a new one to deal specifically with AI, the WTO can map the various policy measures that impact trade and AI to support members in overcoming regulatory fragmentation.

2. Host educational sessions: The WTO Secretariat can conduct a comprehensive assessment of how the current trading system applies to AI. By calling upon members to present their legislation, the WTO can provide constructive input to governments and businesses on how AI usage can facilitate rather than impede trade. Additionally, the WTO can host educational sessions to share its findings and inform AI policy actors and traders.

3. Develop transparency and intellectual property standards: A review of IP rules to better understand how TRIPS and other treaties apply to AI would provide clarity around copyrighted materials. Moreover, by developing best practices and guidance around transparency, disclosure and technical standards, the WTO could support public and private stakeholders in addressing risks such as misinformation and cybercrime associated with AI.

Establishing common ground rules is crucial to ensure fair and safe AI usage in the trade realm. The WTO offers perhaps the only platform wherein this can be done in a truly multilateral way.

This article was written based on insights from the white paper ChatWTO: An Analysis of Generative Artificial Intelligence and International Trade by Penelope Naas and Usman Ahmed.