show index hide index
As artificial intelligence gains considerable momentum in our societies, the European Union’s AI Act positions itself as an ambitious legal framework in the face of the very distinct approaches adopted by China and the United States. This diversity in regulatory methods reveals opposing visions regarding the place of AI in everyday life and the need to protect fundamental rights, while addressing the challenges of innovation and social control. The comparison of European, Chinese, and American models raises crucial questions about the future of this technology with its global implications. The EU’s AI Act and the World: A Comparative Analysis with China and the United States In a world where artificial intelligence is booming, its regulations are evolving in parallel to adapt to growing challenges. The European Union, with its AI Act, represents a pioneering approach focused on the protection of fundamental rights. Conversely,China favors centralized control to preserve social order, while theUnited States adopts a more flexible, innovation-driven approach. This article provides a comparative analysis of the different visions of AI regulation worldwide.AI Regulation in Europe: The AI Act The European Union is leading the way with the AI Act , the first comprehensive legal framework on artificial intelligence. This regulation was designed to address the AI phenomenon comprehensively, encompassing various sectors such as healthcare,justice , education, and even
transportation
This framework imposes strict rules to prevent risks before they materialize, thus ensuring the protection of fundamental rights. The various risk classifications imposed by the AI Act range from a total ban on the most dangerous uses to relaxed standards for low-risk applications. AI through the prism of Chinese valuesIn stark contrast to the European approach, China has opted for AI regulation that focuses on defending the political system and preserving social order. Since 2021, Chinese regulations have emphasized respect for « core socialist values, » prohibiting any use of AI that could undermine national unity. This translates into rigorous security assessments and centralized oversight within the Cyberspace Administration of China (CAC), whose requirements are not necessarily aligned with citizens’ rights. A Fragmented Approach in the United States The United States, for its part, has not yet adopted a federal regulatory framework for artificial intelligence, reflecting a diversity of approaches across states. Thus, the country advocates for fragmented regulation, relying on presidential executive orders and voluntary standards. Under the Biden administration, initiatives such as the AI Safety Institute have emerged, but the primacy of innovation remains a key concern, often prioritized at the expense of a robust legal framework. As a result, regulation varies greatly from state to state, with specific regulations such as those adopted by Colorado against discrimination by AI.
A Fragmented Global Legal Landscape
Thus, the lack of a global consensus on AI regulation complicates the picture, and three distinct international models emerge. The EU’sAI Act, China’s centralized approach, and the United States’ fragmented regulation reveal very different priorities and concerns. In Europe, the emphasis on fundamental rights protection contrasts sharply with China’s more authoritarian control, while the United States’ regulatory flexibility raises questions about risk management. Transparency and Accountability: Three Visions Transparencyalso takes divergent forms in these three systems. In Europe, the AI Act imposes rigorous documentation and traceability requirements for high-risk systems, thus ensuring the verifiability of automated decisions. In contrast, China favors a more opaque approach, where companies are accountable to the state without providing public access to this information. In the United States, transparency remains a nascent concern, dependent on general consumer protection laws, with few requirements at the federal level.It is undeniable that the regulation of artificial intelligence cannot be reduced to a single model. The European Union stands out for its proactive approach and its efforts to protect citizens’ rights, but it faces significant challenges. China, for its part, is asserting itself with authoritarian regulation that tightly regulates the use of AI, while the United States, with its decentralized approach, cultivates innovation at the risk of neglecting security and user rights. This patchwork of regulations shapes our interactions with artificial intelligence and its impact on our societies.
