Artificial intelligence (AI) has become a major strategic issue, influencing global economic and political dynamics. The approaches adopted for its development and regulation vary by country. We can see the techno-liberal model in the United States, the agile model in regions like the Persian Gulf and Singapore, the planner model in China and the regulatory model in Europe. These models are distinguished by their management of innovation, their technological sovereignty and their regulatory control, thus impacting geopolitical balances.
The techno-liberal model in the United States
In the United States, the techno-liberal approach favors rapid innovation and market dominance by major private players such as GAFAM (Google, Apple, Facebook, Amazon, Microsoft). This strategy relies on minimal regulation to promote competitiveness and technological advancement. Companies enjoy great freedom to design and deploy AI solutions, which has led to significant progress and a dominant position on the world stage.
However, this freedom comes with risks, particularly regarding data protection, algorithmic bias and social impacts. The publication in 2023 by the White House of an executive order entitled “Safe, Secure, and Trustworthy Artificial Intelligence” aims to frame the development of AI. They emphasize security, privacy protection and fairness. This order reflects a growing awareness of the challenges posed by limited regulation and the need to establish standards to ensure responsible use of AI. Since its signing, the United States has not issued any new specific federal orders concerning artificial intelligence (AI). However, legislative initiatives have been undertaken at the state level.
On March 21, 2024, Tennessee enacted the ELVIS Act, specifically targeting audio deepfakes and voice cloning. It is the first in the United States to protect artists’ rights against unauthorized use of their voice and image.
On March 13, 2024, Utah Governor Spencer Cox signed bill S.B. 149, known as the “Artificial Intelligence Policy Act.” Effective May 1, 2024, this legislation establishes responsibilities, particularly for companies that do not inform about the use of generative AI. Or when users commit criminal offenses by using AI.
Although no new federal order has been issued since October 2023, the Biden administration has taken steps to regulate the use of AI. In 2023, it obtained voluntary commitments from major companies to govern the risks associated with AI.
The agile model in the Persian Gulf and Singapore
Regions such as the Persian Gulf and Singapore have chosen an agile model, characterized by regulatory flexibility and adaptation to technological developments. These countries are investing heavily in AI to diversify their economies and strengthen their competitiveness. They favor experimentation and rapid adoption of new technologies.
This agility can raise questions about data protection, ethics and social impact, requiring increased vigilance to balance innovation and societal responsibilities.
The planning model in China
China adopts a planning approach, with strong state involvement, in the development of AI. The Chinese government has defined ambitious strategic plans to become the world leader in AI by 2030, supported by massive investments and proactive regulation. This strategy aims to integrate AI into all aspects of society, thus strengthening the country’s technological sovereignty.
This centralization allows for effective coordination of efforts and rapid implementation of technologies. However, it raises concerns about surveillance, social control and respect for individual rights. China has recently announced the implementation of more than fifty new standards for AI by 2026, reflecting its desire to closely regulate the sector while pursuing its ambitions for global leadership.
The regulatory model in Europe
The European Union adopts a regulatory approach, focused on the protection of fundamental rights, ethics and security. This strategy aims to establish a framework of trust for the development and use of AI, while preserving European values. The EU has thus established the AI Act. The objective is to position Europe as a global center of excellence in AI.
This strict regulation can be perceived as a brake on innovation by some industry players, particularly startups, who fear excessive constraints. However, it also offers an opportunity for differentiation on the international stage, by promoting responsible and ethical AI. The EU aspires for its standards to serve as a global reference, thus influencing practices beyond its borders.
The European Union’s Artificial Intelligence Regulation (AI Act) came into force on August 1, 2024. Its main objectives are: protecting fundamental rights in consistency with GDPR, supporting the development of AI technologies while ensuring ethical and secure implementation. Finally, the goal is to create a homogeneous regulatory framework applied uniformly in all member states.
It adopts a risk-based approach, classifying AI systems into four categories:
- Unacceptable risk: prohibited systems, such as those manipulating human behavior or using real-time biometric recognition in public spaces.
- High risk: those subject to strict obligations regarding transparency, data quality and human oversight, particularly in the fields of health, education or critical infrastructure.
- Limited risk: those with transparency obligations, informing users that they are interacting with AI.
- Minimal risk: unregulated systems, such as spam filters or video games.
For high-risk systems, providers must conduct conformity assessments before placing on the market and throughout the system’s lifecycle. They must provide clear information to users about how their software works and its limitations. Finally, they are obliged to ensure appropriate human supervision to mitigate risks.
The regulation provides for fines of up to 7% of the company’s annual global turnover. Fixed amounts can be applied ranging from 7.5 to 35 million euros, depending on the severity of the violation.
Although the AI Act aims to promote ethical and responsible AI, some criticisms emerge, particularly regarding compliance costs for small companies and startups, which could hinder innovation in Europe. The AI Act nevertheless represents a major step towards harmonized regulation of artificial intelligence within the European Union, seeking to balance technological innovation and protection of fundamental rights.
Impact on global economic and political dynamics
These different models significantly influence global economic and political dynamics. The American techno-liberal model favors domination by private companies in the global AI market, reinforcing the economic influence of the United States. The agile model adopted by countries like Singapore allows rapid adaptation to technological developments, attracting investments and talent, and positioning these countries as regional technology hubs. The Chinese planning model aims to establish technological sovereignty, reducing dependence on foreign technologies and increasing China’s influence on the international stage. Finally, the European regulatory model seeks to impose ethical standards, which can serve as an international reference and offer an alternative to more liberal or authoritarian approaches.
Importance of international collaboration
Despite these divergent approaches, the transnational nature of AI makes international collaboration essential. The challenges posed by AI, such as ethical questions, data protection, cybersecurity and socio-economic impacts, require cooperation beyond national borders. Initiatives such as “the global summit on AI risks,” which brought together 28 states in November 2023, testify to the recognition of this necessity. However, in 2023 at Bletchley Park, major players such as Google DeepMind, Anthropic and OpenAI were present, but their commercial rivalry limited any real collaboration. The United States and China, also present, maintained their divergent positions without genuine dialogue, while the European Union promoted regulation focused on security and transparency. Despite initiated discussions, geopolitical and competitive tensions predominated, hindering any progress toward global AI governance. The next AI summit is scheduled in Paris on February 10 and 11, 2025. France, as host, has expressed its commitment to continuing discussions on AI governance. It emphasizes a balanced approach that incorporates both security risks and social and environmental impacts.
Moreover, initiatives such as the call for expression of interest “AI for efficiency” have been launched. The idea is to collect concrete examples of successful AI integrations, in order to strengthen confidence and motivate other organizations to adopt this technology.
