Body
AI is going to fundamentally change world trade. For this, if for no other reason, consideration must be given to updating the organization that oversees the system that governs world trade, the World Trade Organization (WTO). There will be major efforts to improve the trading systems rules and procedures to take into account the vast changes that are likely with the introduction of AI. Those changes in the administration of the agreements dealing with standards (TBT and SPS), the Trade Related Intellectual Property Agreement (TRIPS), and the services agreement (GATS) will be dealt with separately as conditions warrant. Taking an introspective look at the operations of the WTO, discussed here is the application of AI to the three core functions of the WTO—transparency, dispute settlement and negotiation.
My baseline for consideration of what changes are needed or appropriate was the system we have, its net value and its values. These are set out in the WTO guide book that I wrote, published in 2023 by Cambridge University Press, entitled “Revitalizing the World Trading System”. It is a good source for use as background for this discussion.
When I set about the task of considering how the WTO might be changed to take into account the advent of AI, I naturally, or more literally, unnaturally, consulted AI itself. Google AI provided its own view of the relationship of the WTO to the world of AI-enabled trade. It told me that “AI presents a transformative force for international trade, offering both significant opportunities and complex challenges.” It proceeded to opine that “The WTO has a crucial role to play in navigating this new landscape by facilitating dialogue, developing appropriate rules, and ensuring that the benefits of AI are shared globally.”
This is a bit general, but before using AI to delve further into its use for modifying the rules of the WTO, a few cautionary notes need to be sounded:
First, AI is rendering an opinion as to the importance of itself, AI, to the WTO. Is it biased? My father once warned me about bias: “Who tells us that we should drink more milk? The milkman?” AI may have built in biases that are worth being sensitive to.
Secondly, it would be unwise to put too much trust into AI. As Ronald Reagan warned us repeatedly: Trust but verify. (Taken from the old Russian proverb (доверяй, но проверяй). AI can provide wrong or misleading information. It comes with no guarantees. In current jargon, it can “hallucinate”.
Third, AI is not yet primarily innovative in a human sense, that is, introducing new ideas; original and creative in thinking, adding to the total of human knowledge, as opposed to innovative as in a product, featuring new methods; advanced and original designs. For the most part, it is a survey or amalgam of existing thought, ranging over wherever the web- scrapers have gone.
These caveats are not to devalue it. It is merely to state that humanity must understand any tool on which it seeks to rely.
The provenance of AI’s statements is often unknown or insufficiently identified. In reproducing the quoted statements from AI, am I unknowingly quoting some person or persons without attribution? AI does not yet stand on the shoulders of all who went before it (assuming it has done the best possible job of surveying prior relevant thought). Instead, it sits rather heavily on the backs of these largely unnamed and potentially unwilling contributors, its choices based on algorithms that may or may not be available to the public. In its favor, it will often assemble existing thought in a highly very manner. Though combining existing sources, it allows material to be viewed in a new light, in unexpected and often helpful ways.
Commentary Type