Rule-Makers, Rule-Takers, and Domestic Regulatory Space: Emerging Trends in AI Governance through Free Trade Agreements

This article analyses how artificial intelligence (AI) relevant provisions in free trade agreements (FTAs) are emerging as instruments of regulatory governance. Using trade-related databases, namely DESTA and TAPED, we examine 16 agreements involving 50 countries that contain AI-relevant rules across five domains: data privacy, cross-border data flows, open government, data sovereignty/location of computing, and domestic data infrastructure. By adapting a structure-location-commitment framework, we develop an Agreement AI Index (AAI Index) to measure the breadth and depth of these provisions over time and across countries. The findings show that coverage converges faster than legal depth. Further, provisions relating to cross-border data flows and anti-localisation are much more and often binding, unlike open government data and domestic infrastructure provisions that are weak and scattered. Results show cross-country variation in the design of such provisions. Some countries ensure inclusion through hard obligations, while others refer to such provisions through softer clauses. As a result, a small group of countries are ‘rule-makers’ and are leading the way with standard templates, unlike most countries that act as “rule-takers”. We link these patterns to debates on AI-related legalisation, deep integration, and club governance to contribute to the literature on regulation and governance under FTAs.