Artificial Intelligence is installing itself at the center of European political and administrative power with almost no public debate. It appears wrapped in seductive words, such as efficiency, simplification and competitiveness. This is precisely why this is a critical moment for democratic scrutiny. Not because there is a conspiracy, but because a new architecture of power is being built with almost no contradiction.
O AI Act was presented as the great European response to the expansion of AI, with regulation based on fundamental rights and a risk logic. But the political tone changed quickly. The priority is no longer just about regulating well, it is about regulating quickly, simplifying and unblocking processes. In this context, proposals such as the EU Inc.which creates a single European legal entity to accelerate companies and the Package Omnibusdesigned to align and simplify legislation in several areas at the same time, without going through governments or national Parliaments.
The problem is that the Omnibus it does not simplify or improve the legislative process, as it accelerates political decisions, compressing technical debate and public scrutiny. When this method is applied to complex areas such as Artificial Intelligence, there is a risk of reducing the space for civic debate, which also means reducing the ability to understand impacts, anticipate risks and protect rights.
But there is an even deeper paradox in current European discourse. At the same time that the European Union increasingly talks about digital sovereignty, a large part of the technological infrastructure, which supports the digital transition, remains dependent on platforms, clouds and AI models developed by large North American technology companies. Political ambition points to strategic autonomy, while technological reality increasingly approaches dependence.
This dependence is not just technical, it is also institutional. As these companies provide infrastructure, services, consultancy and Artificial Intelligence models for Public Administrations, they inevitably come closer to the centers of political and administrative decision-making. In addition to the formal capture of political power, the big tech currently concentrate technical knowledge, data and technological solutions outside the control of public institutions.
The crossover between AI Act, Omnibus e EU-Inc. reveals a coherent movement. Europe regulates, but at the same time it moves part of this regulation to technical and less visible areas, with codes of practice, harmonized standards and administrative decisions that almost always escape public debate. In these spaces, the influence of large technology companies and consultancies, disguised as lobbyis inevitably greater than that of civil society.
Portugal is adopting this process of transferring power in the current State reform model, based on digitalization and Artificial Intelligence. The promise is to have faster services, less bureaucracy and more efficient decisions, however each automated system and each technological platform also redefines who controls the State’s processes. When technical knowledge is outside the Public Administration, efficiency can be accompanied by dependence and opacity.
None of this is illegal, it simply works through the asymmetry of knowledge, resources and access to decision-making processes. Those who simultaneously master technology, regulation and public procurement always have an advantage. Therefore, the real debate is not between modernization and immobilism, but between two models of power.
On the one hand, technology can be used to improve the functioning of the State. It can speed up processes, reduce unnecessary roadblocks and reinforce transparent decisions, clear rules, equal treatment and democratic control. In other words, it can accelerate good bureaucracy, one that protects the public interest.
On the other hand, what is often happening is not just simplification, but a deeper change. As decisions become dependent on technical systems, platforms and external infrastructures, they are no longer fully visible and controllable. The result is a silent transfer of power, which leaves the public space and enters technical circuits that are less transparent and more difficult to scrutinize.
In this scenario, the so-called digital sovereignty runs the risk of remaining in discourse. Because, essentially, power ends up in the hands of whoever controls the code, data and infrastructure.
When the code governs in silence and the law rushes to accompany it, civil society cannot simply watch. It must demand more transparency and more democratic control before efficiency replaces democracy itself.
Leave a Reply