Chapter 1. Overview of EU AI Law
I. What and How Does EU AI Law Regulate? / 2
II. Overview of EU AI Law / 7

Chapter 2. Scope of EU AI Law
I. Material Applications of EU AI / 40
II. Human Applications of EU AI / 62
III. Territorial Application of EU AI / 85

Chapter 3. Concept of AI System
I. AI / 96
II. AI Systems / 98

Chapter 4. Types of AI Systems and Models
I.
Prohibited AI Practices / 139
II.
High-Risk AI Systems / 193
III.
Limited and Minimally Risky AI Systems / 242
IV.
General-purpose AI model / 245

Chapter 5. AI Understanding

Chapter 6. Obligations of AI System and Model Providers
I.
Requirements for High-Risk AI Systems / 276
II.
Obligations of High-Risk AI System Providers / 327
III.
Obligations of AI System Providers Subject to Transparency Requirements / 389
IV.
Obligations of General-Purpose AI Model Providers / 398
V.
Obligations of Providers of General-Purpose AI Models with Systemic Risk / 418

Chapter 7.
authorized agent
I.
Authorized Representative of High-Risk AI System Provider / 428
II.
Authorized Representative of a General Purpose AI Model Provider / 432

Chapter 8. Scope of EU AI Law
I.
Obligations of High-Risk AI System Operators / 439
II.
Obligations of AI System Operators Subject to Transparency Requirements / 460

References / 467
Abbreviations for laws/treaties / 472

introduction

Artificial Intelligence (AI) is a representative technology with universal applicability across various technological fields, and possesses a dual-use nature, enabling its application in both civilian and military settings. AI technology is rapidly evolving, approaching a level of intelligence rivaling human intelligence. However, concerns about its potential risks are growing alongside its potential.

Against this backdrop, the European Union's (EU) AI law, enacted as the first comprehensive AI regulation law in the international community, aims primarily to protect human health, safety, and fundamental rights from the risks that AI may pose.
Based on product safety regulations, this law is the product of a fiercely contested process involving key EU institutions, 27 member states, academia, industry, and civil society, in a process of seeking to balance technological advancement and regulation.
As a result, this law is having a profound impact on legislation in several countries, including the Act on the Promotion of Artificial Intelligence and Creation of a Trust Foundation, enacted in Korea in 2024, and will serve as an important reference point for future regulatory discussions in the international community.

Discussions about AI in Korea, which aims to become one of the world's top three AI powers, are primarily focused on industrial development, and a negative view of AI regulation seems to prevail among industry, government, and experts alike.
Nevertheless, as a tool meant to benefit humanity, we must never neglect protecting our health, safety, and fundamental rights from the potential harms of AI technology.
Of course, striking a balance between regulation and development is a very difficult policy task, as safety-focused regulations can easily hinder technological development.
However, the minimum necessary regulations must be accepted from the early stages of AI technology development.
This book aims to provide a meaningful reference for domestic discussions and developments in AI regulation by analyzing in detail the complex and sophisticated regulatory framework of EU AI law.


The writing process for this book presented a new challenge for the author. He faced numerous challenges due to his lack of specialized understanding of AI technology and his unfamiliarity with a regulatory approach focused on product safety.
However, through heated discussions with students at Korea University Law School and graduate school lectures, I was able to broaden my understanding by encountering diverse perspectives on AI regulation.
Furthermore, my research on cybersecurity and personal information protection, conducted over the past decade, has been a valuable asset in the writing of this book. AI processes massive amounts of data, especially sensitive personal information, and cybersecurity is an inevitable element of AI regulation, so my experience in these two areas was invaluable.
Additionally, during the writing process of this book, I received assistance from various AI services.
Although there were confusion and limitations in the process of finding precise regulations, AI services served as meaningful companions that expanded the author's thinking about AI and corrected misunderstandings.

Our lives are a series of beginnings and endings, and it was a great fortune to have been able to study EU AI law and write this book at various points in time, especially after March 2024.
In the same vein, I would like to express my sincere gratitude to Director Jeong Cheol of the Korea Economic Research Institute, Korea Economic Association, for allowing me to contribute to the special series "Trends in International AI Norms and Implications" in the second half of 2024.
I would also like to express my gratitude to Director Jo Seong-ho of Park Young-sa, who willingly took on the task of publishing this book in an increasingly challenging publishing environment, and to Ms. Choi Yu-ra, who took on the difficult task of editing it.
I would also like to take this opportunity to express my gratitude to my family, for whom I am always sorry and thankful.

We hope that this book will provide the fundamental perspective necessary to understand the present and future of AI regulation, and make a small contribution to related discussions in Korea and the international community.

October 2025
Park No-hyung