Abstract (En) :
Artificial intelligence is probably the most important vector of security and law enforcement transformation. The need of security is a response to the defense of fundamental interests of states. How to protect these interest with artificial intelligence tools and technologies from foreign states? Beyond the major state powers, it is necessary to address the economic environment of artificial intelligence. The main players in the development of artificial intelligence in the world are the largest technology companies : GAFAM and BHATHX. The reality is that their capacity to influence extends beyond the borders of the United States or China and their power is not limited to companies, but extends to States, NGOs, international and regional organizations. This hegemony is a reality also in Europe.This creates a situation of dependency. We have chosen to work on the basis of the theory of dependence that emerged at the end of colonisation, in an atypical approach that opens up the field for future research.

Keywords : Artificial intelligence, security, fundamental interests, dependency

INTRODUCTION

In 1996, in an article entitled « The World Wide Web as Super-Brain: from Metaphor to Model », Francis Heylighen and Johan Bollen of the Vrij Universiteit Brussel (VUB) put forward some interesting proposals concerning the development of the « super-brain » and what enables it to learn, without omitting to point out that it is not the brain itself that thinks, but the users of the web. Indeed, the power of this ‘super-brain’ lies in the tenuous link with its users, a self-referential link. Algorithms have also developed, which (by analogy with the human brain) strengthen the links and weakens those that are less frequently used. Using the principle of transitivity, the construction of new links can be automated. But none of this means that this « super-brain » can actually think independently of the users that make it up (Mortier, 2019). Today, the conditions are ripe for a widespread evolution of AI technology techniques: availability and diversity of data, development of offers and performance of IT devices and equipments (Marellin, 2021).

In other words, AI, which is both a discipline and a scientific modeling tool, brings together all the modeling representation techniques that allow to simulate a phenomenon or a situation by computer. This phenomenon can be physical, chemical, biological, medical, sociological, financial or demographic. From these simulations, answers to questions, elements of understanding of the phenomenon in question and even components of prediction were learned. In practice, one or more algorithms implemented within a computer program run on one or more computer processors. An algorithm is a hierarchical set of logical operations to execute in order to solve a problem or answer a question (Jean, 2021).

The ability of nations to develop AI firms is critical to their competitiveness. Such firms provide tools and services to the growing number of companies adopting AI. Indeed, the global percentage of large companies using AI in at least one function or business unit increased from 47 percent to 58 percent from 2018 to 2019 (Castro & McLaughlin, 2021). In 2023, 83% of companies say AI is their priority for the coming years1. Insofar as power is exercised on an informational basis, the determinants of success or advantage have less to do with the information we possess and the knowledge we can deduce from it for strategic purposes, than with the management of increasingly automated information (Atif et al., 2022). In this sense, AI is a power issue.

The Artificial Intelligence (AI) has giant possibilities to optimize the fight against crime and strengthen national security. In the conditions of unimaginable accumulation of information and the need for rapid decision-making, only the use of AI can lead to success. Intelligence, counterintelligence, forensic science, counteracting organized crime, rapid processing of available information, drafting of varied decisions, creating plans and multivariate scenarios, performing various analyzes is a time-consuming process. Only its use can significantly shorten this time and thus dramatically increase the possibilities for detection, prevention and curbing crimes (Radulov, 2019). In brief, Artificial Intelligence (AI) and Machine Learning (ML) can be a great cyber defense strategy, it can also be a double edged sword (Meghani, Essomba, Chrzanovski, 2023). AI belongs to the new class of technological objects which, as simple objects of value, portable, transferable, easily convertible, present themselves as the ideal object for a global hyper-capitalism, supported by flows of circulation of value and its unprecedented capacity to create value, AI is inevitably and completely taken over by geopolitical forces (Atif et al., 2022). That’s the reason why AI is and will increasingly be indispensable for national security. It is inconceivable to exclude such technology from security strategies. More and more, the regalian services are using artificial intelligence tools, both for judicial investigations and for intelligence or administrative police.
1 https://lesmakers.fr/statistiques-intelligence-artificielle/

Clearly, the powers that be are now engaged in an AI race, although only two countries look set to win it in the near future. Only two countries seem likely to win it in the near future: they are, of course, the United States and China (Thibout, 2018). As for Europe, it is trying to regain a leading position among other players such as Russia and Korea (Atif et al., 2022). The risk is not in the use of artificial intelligence but in the algorithms or tools that enable its use. Most of the algorithms and tools are developed outside the European Union and place the Member States in a certain form of dependence on foreign powers. This of course generates risks for national security and the protection of the fundamental interests of European states.

This situation of dependence is reminiscent of the theories developed in the context of decolonization. In particular, the theory of dependence based on the work of Theotonio Dos Santos (1970). This theory is a critique of structuralism, the theoretical corpus developed within the framework of the United Nations Economic Commission for Latin America (CEPAL) in the 50s and 60s. However, this theory is of particular interest in the context of artificial intelligence dependencies.
In a singular way, we will attempt to analyze the European situation through the prism of this theoretical framework. There is no intention here of considering Europe as a developing geographical zone, but rather to demonstrate that in certain sectors of activity, the old Western powers can find themselves in a situation of dependence, with all the consequences that this can produce in society asa whole.

While we will live in aworld of permanent low-level conflict, often unnoticed, undeclared and unending, and one in which even our allies may also be our competitors (Galleoti, 2022), after recalling the context in which artificial intelligence is evolving and becoming unavoidable, we will address the risks to which European countries are currently exposed. But first we will look at the theory of dependency, whose historical context is reminiscent of the current situation of European states in the AI race…

Stéphane MORTIER,
Associate lecturer Université Gustave Eiffel (France), Research Centre of the French
National Gendarmerie (CREOGN)