Editors’ Introduction

Welcome to the third edition of the Law Over Borders Comparative Guide to Artificial Intelligence. For this edition I am delighted to welcome global co-editor Dean Harvey of Ashurst Perkins Coie, who brings a wealth of knowledge and a valuable North American perspective to the topic of AI law and regulation. It has become increasingly apparent (and I am sure no surprise to you, the reader) that the phrase “many hands make light work” is apposite in describing the burgeoning demands of editing a global guide on AI. As ever we are indebted to the country contributors who have sacrificed time and effort to provide local insights on the ways in which artificial intelligence is regulated across the globe.

The second edition of this guide was published in July 2024. In the two years since we have seen developments in the scope and utility of AI models and systems, which typically have opened more novel issues of liability. The topic on everybody’s lips is “Agentic AI”: put simply, AI that can exercise judgment and do things for us (in contrast to the current crop of generative AI models that respond passively to prompts). Agentic AI opens new horizons in terms of AI capability, but with that inevitably comes execution risk. The principles of modern human agency law have evolved over the last four hundred years, yet no such equivalent exists in relation to the actions of AI agents. These are horizons we expect to be resolved in future editions.

In a world which is increasingly dominated by global influence and hegemony, concerns have been raised at national level over the cooling effects of AI regulation. These worries were first articulated in the EU by former European Central Bank president Mario Draghi in his substantive 2025 report on European competitiveness. Ultimately this led to the 2026 EU Digital Simplification initiative, one impact of which has been to defer the commencement of the High-Risk AI System regime in the EU AI Act. In the US, a federal policy of limiting AI regulation to preserve AI innovation and competitiveness nationally has led to a pre-emption battle between State and Federal governments over what has been termed a patchwork of inhibitory state AI laws. Finally, the UK, caught perhaps between these two major economies, has sought to reinforce its own pre-eminent AI position by shelving a much-touted law on frontier AI models.

In this climate, the increasing dominance of US-led hyperscalers has also fostered nervousness over sovereign AI capabilities, with many countries now considering how to reinforce their own AI infrastructure, reduce US dependencies and channel more investment into local data centre development.

Finally, in this vein, the extent to which national regulators, such as those in the EU and the UK, can influence and restrain US businesses is becoming an increasingly divisive issue — particularly in light of the tension between US constitutional free speech principles and the desire of those regulators to prevent content deemed harmful (such as deepfake nudes) from being created and distributed.

Away from macroeconomic influences, statute and case law relating to AI continue to evolve and develop, although many of the more intractable issues — such as the tension between intellectual property content rights holders and AI developers over the legitimacy of web-scraping — remain to be resolved definitively. The country Q&A chapters in this guide provide valuable insight into how such matters are being resolved nationally.

This third edition of our Artificial Intelligence Law Over Borders Comparative Guide endeavours to maintain the pace on what continues to be an exponential evolution of global AI law.