Artificial Intelligence (AI) poses a unique challenge to the international community, demanding a flexible approach to governance. Its rapid advancement, increasing decision-making capabilities, and significant impact necessitate regulation aligned with human principles: oversight, understanding, and ethical reasoning. The primary impetus for legal intervention is AI's unpredictable nature and the numerous identified risks, including behavioral manipulation, exploitation of vulnerabilities, and threats to citizen trustworthiness. However, regulations must not stifle innovation. While national and local regulations are crucial, a global framework is necessary to overcome transnational limitations. A universal regime could offer model laws, drawing on global expertise, as a first step towards greater harmonization. The disparity in AI investment and technological capabilities creates a risk of marginalizing developing and Least Developed Countries (LDCs), potentially leading to conflicting interests among nations. This necessitates a new global AI regulatory authority to address current and emerging challenges, anticipating rapid technological developments, balancing stakeholder interests, and providing the necessary expertise and resources for effective regulation.

The regulation of AI has been a subject of discussion for decades, with nations and supra-national institutions competing for leadership in establishing regulatory frameworks. Questions about the most effective regulatory approach and the necessity of law remain. Lawmakers must possess a thorough understanding of AI to establish appropriate rules, as traditional legal frameworks are often challenged by AI's unique characteristics. The interplay between law and AI is not new, drawing parallels to previous regulatory efforts concerning technology. The role of international law in AI's development is heavily debated, encompassing its fundamental principles, ownership, and liability. AI rules are designed to safeguard public interests, ethics, and existing inequalities. Obstacles to effective regulation stem from the lack of consensus on AI's definition, the various applicable regulatory theories, the breadth of AI's impact across various sectors, and the ongoing regulatory race amongst different legal regimes. Existing international norms and rules, such as those within international trade law, criminal law, economic law, human rights law, humanitarian law, intellectual property rights, and regional regulations like the EU AI Act, represent the initial legal response to AI. Scholars debate the efficacy of regulation, its impact on various stakeholders, and the optimal balance between strict and soft rules. The discussion encompasses a wide range of challenges, including the constant evolution of AI, lack of transparency, difficulty in addressing accountability, and the differing viewpoints on whether more or less regulation is needed. The EU AI Act, the Council of Europe's framework convention, UNESCO's recommendations on the ethics of AI, and US federal and state regulations represent some of the significant current regulatory initiatives.

The article uses a qualitative research approach, employing a comprehensive literature review to analyze the theoretical framework for global AI governance. The analysis centers on the role of a hypothetical global AI regulatory authority and the intricate balance required to manage the participation of diverse stakeholders: states, individuals, civil society organizations, corporations, and international organizations. The authors examine the existing literature on international law, AI regulation, and global governance to develop a conceptual model illustrating the interactions between these stakeholders and the regulatory authority. They analyze the functions of this authority, including establishing international AI law and monitoring its enforcement domestically. The study draws extensively from various legal instruments, reports, publications, and academic work to inform the discussion. The review examines national and international legislative efforts, the roles of diverse stakeholders in AI governance, and challenges to effective regulation. The authors do not present any original empirical data or conduct original field research, focusing instead on the secondary literature and analysis of existing frameworks and proposals for global AI governance.

The article emphasizes that establishing a global AI regulatory framework is exceptionally complex, requiring significant effort to balance diverse stakeholder interests. States play a crucial rule-making role, requiring their consent to establish and enforce AI rules domestically. The authors propose a conceptual model (Figure 1) illustrating an AI regulatory authority responsible for rule-making, enforcement, and oversight, with representation from states, international organizations, private companies, and civil society. The article highlights the difficulties states face in reaching consensus on AI regulation due to conflicting national interests and the challenges associated with international law-making. The slow pace of developing international AI regulation is emphasized. While national and regional initiatives like the EU AI Act and the Council of Europe's framework convention demonstrate some progress, achieving global coordination remains a significant obstacle. Different approaches to AI regulation exist, with varying preferences for hard or soft law mechanisms. The authors discuss the roles of various actors, including states (as primary rule-makers), natural persons (affected by AI systems), civil society organizations (advocating for ethical AI), corporations (developing and implementing AI), and international organizations (facilitating cooperation). They note the limitations of international organizations in enforcing AI regulations domestically, emphasizing the crucial role of state compliance. The review of current AI regulations from the EU, the Council of Europe, China, and the US highlights the diversity of approaches and challenges in harmonizing global standards. The article argues that a global AI regulatory authority should leverage both hard and soft law instruments, starting with non-binding guidelines and standards before transitioning to more legally binding rules. Effective monitoring and enforcement of AI regulations are also crucial, requiring cooperation and capacity building, particularly for developing nations. The challenges of monitoring AI activities given their complexity are also noted.

The findings suggest that the establishment of a global AI regulatory authority is a complex but potentially necessary step for effective global AI governance. The slow progress in international AI law-making underscores the difficulties of balancing diverse national interests and the need for a flexible approach that accounts for the fast-paced technological developments. The authors' model of a global regulatory authority, while hypothetical, offers a framework for considering how this complex interaction of interests could be managed, including the different roles of various stakeholders and the tension between hard and soft law. The success of a global framework will depend significantly on states' willingness to cooperate and compromise. The article's analysis of existing national and international regulatory initiatives highlights the need for a more coordinated and comprehensive global strategy. Further research is needed to investigate potential models and their efficacy in managing the complex technological, economic, and political dynamics surrounding AI.

This article demonstrates the profound complexity of global AI governance and the challenges of harmonizing diverse national interests in this rapidly evolving field. The proposed model of a global regulatory authority, while aspirational, provides a framework for understanding the intricate interactions among key stakeholders. The progress made through initiatives like the EU AI Act and the Council of Europe's framework convention offers some hope, but achieving comprehensive global regulation will demand continued cooperation and a pragmatic approach that balances innovation with the need to mitigate risks. Further research on the optimal structure and functions of a global AI regulatory authority is strongly recommended.

The study primarily relies on a review of existing literature and does not include original empirical data or field research. The analysis primarily focuses on Western regulatory initiatives, potentially neglecting perspectives from the Global South. The authors acknowledge that the hypothetical model of a global regulatory authority is subject to various potential future modifications based on evolving technological advancements, political dynamics, and the evolving international legal landscape. The paper primarily focuses on the legal aspects of AI regulation, potentially neglecting the interplay with political, economic, and social factors.