THE GOVERNANCE OF SILICON: GEOPOLITICS OF CODES AND THE ASYMMETRY OF DIGITAL POWER

Translated by Benedetta Calice

The supremacy of power

We are accustomed to viewing the law as the pinnacle of civilisation, the rational barrier that advanced societies erect against the brutality of the state of nature. This is the typical perspective error of post-historical powers, foremost among them the European ones, who are convinced that norms can shape reality and that a well-drafted treaty can alter the balance of power. But geopolitics, which feeds on deep-seated and anthropological factors, disregards good intentions. In the most heated fault line of our contemporary world — the struggle for dominance over Artificial Intelligence — the law is not the end, but the weapon with which large human agglomerations attempt to protect or expand their sovereignty.

Today, the concept of power is expressed in terms of computing power, advanced semiconductors and vast amounts of data. Those who possess these assets determine the psychological and social stance of populations. Those who lack them attempt to compensate through excessive regulation. This conflict is not about technology, but about the survival of peoples throughout history.

The three stances of the empire: Washington, Beijing, Brussels

The geography of digital law closely mirrors the geopolitical ambitions of the three major blocs. Laws do not arise in a vacuum, they reflect the collective unconscious of the nations that create them.

1. The American model: the expansion of private hegemony

For the United States, technological innovation is an extension of the Frontier doctrine: a partially anarchic space where the private sector preempts the state, accumulating enough power to become an empire. The stance of the American legislature has historically been one of non-intervention. Washington intervenes only when the concentration of power threatens national security or the established order.

Macroeconomic data confirms this gap: according to reports on global AI investment, private investments in the United States reached $67.2 billion in 2023 alone, nearly nine times the total amount allocated by the European Union to the same sector.

U.S. law evolves ex post facto, through court precedents rather than through codified statutes. The landmark Authors Guild v. Google (2015) ruling, which established the concept of fair use for the digitization of books, serves as a psychological pillar for today’s Big Tech companies: training models on copyrighted data is considered lawful as long as it produces “transformative” value. The executive order signed by the White House in October 2023 on AI safety does not contradict this approach: it imposes stringent controls only on systems that exceed the computational threshold of $10^{26}$ floating-point operations (FLOPS), a limit that preserves ordinary commercial innovation while leaving intact the dominance of American companies over 70% of the global market for language models.

2. The Chinese model: technology at the service of dynastic harmony

In Beijing, the law is not designed to protect the individual from the sovereign, but to protect the state from the seismic upheavals of society. Artificial Intelligence is perceived by the leadership of the Chinese Communist Party as an extraordinary tool for social stabilisation, but also as a potential vehicle for Western ideological infiltration. The legal response of the People's Republic has been surgical.

The provisions of the Algorithmic Recommendation Regulations (2022) and the subsequent regulations on Generative AI (2023) stipulate that machine-generated content must “reflect fundamental socialist values” and must not in any way “subvert state power”.

Beijing has realised that social control hinges on technology: its domestic surveillance system, supported by over 200 million facial recognition cameras linked to predictive models, serves to prevent the breakdown of social cohesion. It is no coincidence that China holds over 60% of the world’s patents related to computer vision technology, a statistic that reflects a clear strategic priority: maintaining internal imperial stability.

3. The European Union’s frustrated ambition: the AI Act

It is against this backdrop that the European Union finds itself: a geopolitical entity lacking a common defence, devoid of global technological leaders (no European company features among the world’s top twenty in terms of market capitalisation in the tech sector), yet convinced that it can exercise moral leadership through written legislation.

The AI Act, finally approved by the European Parliament on 13 March 2024 and fully operational in accordance with its phased timetable, is the manifesto of this illusion.

The European Union operates according to the logic of the so-called ‘Brussels Effect’, as theorised by legal scholar Anu Bradford: the idea that the strictness of the European regulatory market forces foreign giants to comply with our standards in order not to lose the EU’s consumer base. This was the case with the GDPR on privacy, which resulted in record fines — recall the €1.2 billion fine imposed on Meta by the Irish Data Protection Commissioner in May 2023 — but which has not in the slightest diminished the Old Continent’s strategic dependence on overseas cloud infrastructure.

The AI Act introduces a risk-based classification of systems:

• unacceptable risk: systems that are outright banned, such as social scoring or indiscriminate scraping of facial images;
• high risk: systems subject to stringent compliance requirements (fundamental rights impact assessment, data traceability);
• sanctions: for the most serious infringements, fines of up to €35 million are envisaged, or up to 7% of the non-compliant company’s annual global turnover.

Society punished by regulatory idealism

It is precisely in this rift that the divide between law and society reveals its deepest contradiction. The European regulatory framework is designed to protect citizens, but the sociological effect risks being diametrically opposed. When ex-ante regulation is so stringent, home-grown innovation is stifled at the outset. An analysis by the Centre for Data Innovation reveals that the compliance costs for a small or medium-sized European enterprise seeking to develop a high-risk AI model can exceed €300,000 in legal and procedural fees alone.

The result is clearly evident in the flow of human capital: Europe’s finest scientists and engineers are migrating to laboratories in Boston, Silicon Valley or Shenzhen, where legal flexibility allows for immediate experimentation. Europe protects the rights of citizens who consume technologies developed by others, whilst failing to produce them itself. It is the reduction of society to a legal 'Indian reservation': protected but sterile.

Furthermore, this conflict has repercussions for the stability of democratic institutions. In the United States, the extreme protection afforded by the First Amendment makes it difficult to remove deepfakes or algorithmic disinformation, exposing the electorate to campaigns of psychological destabilisation orchestrated by hostile external actors. In Europe, by contrast, attempts to regulate online content through the Digital Services Act (DSA) leave the door open to accusations of creeping state censorship, undermining the bond of trust between citizens and institutions.

The ‘Digital Westphalia’ that will never be

European governments are under the illusion that the world is moving towards a gradual legal convergence, a sort of global constitutionalism for the digital sphere. This interpretation ignores the basic principles of international relations. There will be no ‘Digital Westphalia’ because there is no shared interest in limiting one’s own cognitive weapons. For the real powers, the predictive effectiveness and military efficiency of AI models are existential matters.

Without the material power to enforce it, the law becomes nothing more than empty rhetoric. Until Europe possesses the industrial and infrastructural sovereignty necessary to uphold its own laws, the AI Act will not be a shield for European society, but an act of capitulation of its geopolitical significance. For in history, the rules are not written by those who codify them on a sheet of paper, but by those who possess the material force to compel others to respect them.

Mondo Internazionale APS - Riproduzione Riservata ® 2026