Panel 2.41

A wave of Executive Orders from the United States on AI proclaim, and seek to secure, American “global AI dominance” – not merely global leadership - as a core value in the short-term. Despite being voted down as part of the One Big Beautiful Bill Act, a moratorium on AI legislation by the U.S. States has been imposed by Executive Order 14365 of 11 December 2025. The prohibition on all AI-related legislation not concerned with child safety or AI infrastructure, seeks to counter the ‘Brussels effect’ on States’ AI regulatory initiatives, given that, at the federal level, the US perceives law and regulation to be anathema to innovation. At the same time, 20 US states have adopted comprehensive data privacy legislation, and several more have similar GDPR-inspired legislation under consideration. The debate has become transatlantic, with the US actively opposing digital regulation in the EU and its Member States. The simplification and delegalisation debates taking place in the EU around the Digital Omnibus, combined with increased use of soft law in EU digitisation law and policy, including external relations, are argued here to be seen as a product of a “Washington effect” and a process of the enhanced presidentialization of the EU system. 

These debates matter on account of the fragility of the key legal instruments of EU-US relations that are significantly eg EU-US Transatlantic Privacy Framework. 

The paper explores, form an institutional and temporal perspective, the question of transatlantic divergence and convergence in law and policy-making with shifts in administration.

This presentation examines the emerging constitutional and regulatory conflict between U.S. member states and the federal government over authority to govern artificial intelligence (AI). Drawing on comprehensive analysis of federal executive orders, state legislation across all fifty states, and regulatory guidance documents from global AI policy databases, this research reveals a deepening fracture in American AI governance. In the absence of comprehensive federal legislation, states have enacted divergent regulatory frameworks addressing algorithmic accountability, bias prevention, and fundamental rights protection. Simultaneously, the federal government has attempted to use executive orders with the aim of constraining state authority over AI governance. This federal-state conflict raises critical questions about the allocation of regulatory power in emerging technology domains and the capacity of fragmented governance structures to protect fundamental rights consistently across jurisdictions. The proposed presentation argues that this regulatory fragmentation reflects broader challenges facing public law in coordinating responses to transformative technologies that transcend traditional jurisdictional boundaries, with significant implications for how democratic societies balance innovation, security, and fundamental rights protection in an increasingly fractured governance landscape.

Digital governance has emerged as a contested domain of global regulation, often portrayed as a strategic battleground among the United States, China, and the European Union. Yet the metaphor of rivalry obscures a more complex dynamic in which competition, imitation, and regulatory diffusion coexist. This paper examines the evolving landscape of AI regulation through an interdisciplinary lens, combining legal analysis with insights from ethics, political science, and psychology. Focusing on the European Union, Brazil, and Japan, it explores how distinct normative commitments, institutional structures, and societal attitudes shape emerging AI governance models.

The EU’s 2024 AI Act, the first comprehensive AI statute worldwide, exemplifies a value‑based, product-oriented approach rooted in the Union’s legal path dependencies and internal institutional dynamics. Brazil’s AI Bill, currently under legislative review, reflects a hybrid trajectory: initially principle-based, later inspired by the EU model, and most recently reoriented toward sectoral enforcement and elements of self-regulation. Japan, by contrast, advances a government-led, soft law framework complemented by targeted sectoral regulation, emphasizing non-binding guidance. These approaches unfold against differing public attitudes toward AI and varying balances between legislative, executive, judicial, and private actors.

The paper argues that AI regulation constitutes a high-stakes arena in which jurisdictions simultaneously compete and converge in pursuit of an as-yet undefined global standard. Collaborative governance models – multidimensional in the EU, state-centered in Brazil, and nonformal and supportive in Japan – illustrate how legal frameworks are increasingly shaped by political, ethical, and psychological factors in defining the rules of the digital future.

This paper examines whether contemporary AI regulation is giving rise to a shared cross-jurisdictional standard for human oversight in high-stakes automated decision-making. Through a comparative analysis of the EU AI Act, recent U.S. executive and agency-level frameworks, and Asian developments such as South Korea’s 2024 AI Framework Act and Japan’s AI Promotion Act, I explore whether these systems reflect converging constitutional values or merely parallel policy instincts.

Across these jurisdictions, the paper identifies several emerging commonalities: (1) explicit rejection of fully automated decision-making in contexts affecting fundamental interests; (2) an insistence on “meaningful” rather than symbolic human involvement; (3) requirements for transparency, traceability, and explainability sufficient to enable genuine oversight; and (4) growing awareness of automation bias and the cognitive limits of human supervisors. These features suggest that human oversight is becoming a structurally necessary safeguard in democratic and technologically advanced regulatory regimes.

However, closer analysis reveals that this alignment is not yet a constitutional convergence. The EU embeds human oversight within a rights-based, legally binding framework; the U.S. approach remains sectoral, fragmented, and largely administrative; and Asian models prioritize safety, societal trust, and human-centric design but rely heavily on soft-law guidance. Enforcement architectures also diverge sharply and still develop.

The paper concludes that what is emerging is not a global constitutional minimum, but a family of regionally grounded oversight models animated by shared concerns about human control. These patterns raise important questions about whether functional equivalence—not formal constitutionalism—may become the basis for future global AI governance.

The recent surge in large language models (LLMs) has led to the popularization of generative AI (GenAI), enabling the development of various chatbots, such as ChatGPT, Gemini, Claude, and others, that can generate new content by predicting the next word in a sentence. As GenAI services become increasingly integrated into daily life, understanding their benefits, potential risks, and societal impacts is crucial. Consequently, many jurisdictions are now grappling with the challenge of regulating this rapidly evolving technology. 

In this paper, I will examine the regulatory approaches to GenAI in China and the United States, the two largest economies and innovation hubs, with a specific emphasis on California. The paper mainly employs comparative legal methods contrasting GenAI frameworks in China and California, such as the PRC Interim Measures for the Management of Generative Artificial Intelligence Services (Interim Measures on GenAI), Governor Newsom’s Executive Order on GenAI (Executive Order N-12-23), California’s Safe and Secure Innovation for Frontier Artificial Intelligence Models Act (SB 1047) and Transparency in Frontier Artificial Intelligence Act (SB 53). My analysis will also include a discussion of the enforcement and implementation strategies in the two jurisdictions, such as public versus private mechanisms, as well as technical alignment techniques. 

Finally, this paper aims to provide a comprehensive and nuanced view of the ways GenAI is being regulated in these two regions and contribute meaningful insights to the ongoing discussion on responsible AI development and implementation, fostering understanding and dialogue across jurisdictions to influence the future of GenAI.

Over the years, a substantial number of new EU legal acts concerning digital markets have come into force, including the Digital Markets Act, the Digital Services Act, the P2B Regulation (hereinafter: P2B) and Directive 2019/2161 'Omnibus'. In an effort to simplify the law and ease the burden on businesses, the proposal for the 2025 Digital Omnibus Regulation calls for the repeal of the P2B.

Based on this, the study's research question is whether far-reaching legislative action is necessary, and whether it reduces the rights of business users. To answer this question, the study will firstly examine the possible overlaps in the aforementioned legal acts concerning obligations relating to the ranking of products/offers. The research focuses on the scope of these acts, their objectives, the definitions of ranking and the nature of the obligations. The scope of the study was narrowed to ranking due to its importance and the fragmented nature of the legal regulations in this area. This is followed by a critical assessment of the arguments for and against repealing P2B, considering any found overlaps.

The study aims to make proposals for interpreting the provisions, in order to avoid overlaps and ensure the coherence of the norms that govern product ranking. Lastly, in the event of P2B being repealed in the future, the study includes proposals for interpretation and enforcement that would protect the rights of business users.

The research's main methodology is a formal-theoretical approach. The study contains analyses of legal acts, proposals and ‘soft law’ documents.