Artificial intelligence in the public administration system of Russia and the European Union: comparative legal analysis. Part II

Full Text

The experience of the European Union The European Union is not only a difficult neighbor and the most important trading partner of the Russian Federation in the recent past, but also a serious player in the emerging new reality. The digital law of this interstate association, which is currently taking shape, sets the vector for the further progressive development of the global digital space [1]. The ideas that are being implemented in European countries are also being implemented in other countries. Of course, the EU’s experience in regulating the use of artificial intelligence cannot but be of interest to a Russian researcher and legislator. It is generally accepted that the legal systems of the EU countries are overwhelmingly included in the so-called continental legal family, which confirms the importance of analyzing such a source of law as a regulatory legal act. It should be recalled that in the EU, the legislative initiative belongs to a greater extent to the European Commission. Moreover, EU law opens opportunities for the European Commission to independently adopt a regulatory legal act, which means that the law-making significance of this public authority is considered extremely high. In general, according to the so-called joint decision-making procedure, the act will be considered approved when a consensus is reached between the European Parliament and the EU Council. It is noteworthy that of all EU public authorities, only the European Commission is working purposefully to prepare legally relevant documents in the field of regulating the use of artificial intelligence. This circumstance can be explained by the fact that the European Commission, as an executive authority, reacts more quickly and sensitively to the changes that the digitalization process brings to European society. The adoption of the Machinery Directive of the European Parliament and the Council of the EU in May 2006, which regulates the use of technological innovations in sufficient detail, can be considered as a starting point in the formation of the digital law of the EU countries. However, the regulation of the use of modern digital technologies in Europe began relatively recently. In this context, the European Commission’s proposal to form an Expert Commission in March 2018, which was to identify priority areas for the upcoming legislative process, became a turning point. It is believed that the creation of the Expert Commission was due not only to the urgent need to form some generally accepted rules in the context of rapidly increasing digitalization, but also to the legislative policy of the EU. Recall that back in December 2017 the European Commission, the Council of the EU and the European Parliament have adopted a joint so-called “Declaration of Legislative Priorities” (Joint Declaration on the EU’s legislative priorities), in which the process of norm-setting in the field of digital relations was given special attention. This document appealed to digital innovation in general, and artificial intelligence was considered in it along with other innovative ideas. Moreover, this document outlined other key areas for the further legislative process, among which are ensuring the safety of EU citizens, guaranteeing their rights and freedoms, improving migration policy, ensuring full employment and economic growth, and the investment policy of EU member states. As for the Expert Commission itself, its goal was to identify those areas “from improving healthcare to safer transport and more sustainable agriculture”, in which legal regulation of the use of artificial intelligence will find its practical application in the near future. At the same time, the European Commission proposed following the so-called principles of European morality as a kind of guideline in order not to undermine the confidence of the EU population in technological progress. As Andrus Ansip, Vice President of the Digital Single Market Commission, later noted: “Step by step, we are creating comfortable conditions for Europe to make the most of what artificial intelligence can offer. Big data, supercomputers, and bold investments are necessary for the development of artificial intelligence, along with broad public discussion combined with ethical principles in its implementation. As it turned out, when promoting the latest technologies, public trust is a prerequisite”1. In April 2018, the Expert Commission published the document “Artificial intelligence in Europe” (hereinafter referred to as “AI in Europe”), which provides a detailed overview of the entire range of issues directly related to the use of not only artificial intelligence, but also other types of artificial intelligence. The document consists of two large sections. The first section is devoted to the current legislation, the second to the description of the most widespread digital technologies in the EU countries. We should also add that “AI in Europe” is not an independent regulatory legal act, its status is designated as a working and complementary document. Nevertheless, its scientific value is quite high, as its provisions reflect the specifics of the European approach to the use of artificial intelligence. Moreover, references to “AI in Europe” are found in several analytical and scientific papers published abroad. A feature of “AI in Europe” can be called a clearly traceable appeal to the protection of consumer or user rights by digital technologies. The document emphasizes that back in 1985, the then-existing EEC approved the so-called “Product Liability Directive”, which protects consumer rights in the retail sector. It is in the context of this Directive that the document we are considering sets out the key provisions regarding the use of artificial intelligence in modern Europe. As for the legislation of individual EU member states, in this case, the authors give an example of Germany as a pioneer country. Thus, the introduction of norms regulating traffic using artificial intelligence seems to be very successful. In this case, we are talking about the Traffic Law (Street Traffic Act), which was approved in 2017. “AI in Europe” also refers to those countries outside the association, where work is also underway to form digital legislation. It states that “many states in the United States are currently considering the need to adopt legislation on autonomous vehicles”2, while emphasizing that these initiatives vary significantly from state to state, as regional legislators set different goals. In addition, the problem of combating cybercrime and protecting personal data is acute in the United States. Japan is cited as an example of the most successful state, where the Ministry of Economy, Trade and Industry initiates a discussion of legal issues related to the recognition of legal responsibility for artificial intelligence. The second chapter of the document provides a rather interesting interpretation of artificial intelligence, which refers to digital technology, the components of which can be material parts or devices (sensors, actuators, hardware), various software components and applications, or “the data itself”. The mechanism consisting of these elements performs the functions of data transmission. It should be noted that this definition is perhaps the only one that we have been able to identify during the analysis of the most significant regulatory legal acts issued in the EU. Moreover, this interpretation differs significantly from the Russian understanding of artificial intelligence, which assumes its autonomy from human volition. In general, the document we are considering differs significantly from Russian legal sources in its structure and method of presentation of key provisions. For example, throughout the text of the document, we are talking about artificial intelligence and other types of digital technologies, and these concepts often replace each other. And only in the final part does the legislator briefly explain the difference between artificial intelligence and the Internet of things. It seems that it is impossible not to pay attention to planning in the digital sphere, even though rulemaking in the field of forecasting, programming and goal setting is not a distinctive feature of the European legal tradition. As a result, unlike the Russian legal system, the concept of strategy has not become widespread in the EU legal field. It should be recalled that in June 2014, Russia adopted the Federal Law “On Strategic Planning in the Russian Federation”, which consolidated the status of a number of such documents at the federal and regional levels. Just like in Russia, the EU periodically adopts strategic industry planning documents, which list tasks and measures implemented for their implementation. The digital industry is no exception. One of the latest such regulatory acts was the “Coordinated Plan on Artificial Intelligence 2021 Review”, approved in April of the same year3. The document outlined tasks: · increasing the investment attractiveness of projects that use artificial intelligence; · coordination of the main provisions of the policy on the use of artificial intelligence with all participating countries; · acceleration in the field of investment activities in technologies related to the use of artificial intelligence. One of the most debated legislative initiatives is the so-called “The European approach to artificial intelligence” (hereinafter referred to as the European Approach), which was approved in stages by the European Commission starting in 20214. It is difficult to call this initiative a regulatory legal act, since, firstly, it is a set of three separate coordinated and mutually complementary documents, and secondly, it is not binding for all EU member states. The introductory part of the initiative states that the European Approach is “based on best practices and public trust, it aims to enhance Europe’s research and industrial potential and ensure the fundamental rights of our citizens”5. The European approach is expected to help create “a sustainable Europe where people and organizations can benefit from artificial intelligence”6. It should be noted that the initiative is also being discussed among Russian researchers, which confirms its importance in the eyes of specialists7. The first document is called the “Communication on fostering a European approach to artificial intelligence” (hereinafter referred to as the Program). In this document, a number of terms are introduced into practical use, which are more declarative in nature and are already widely used by European analysts. For example, such a concept as the “digital decade” is quite common, which implies a period of “active innovation in the life of European society”. At the same time, the Program emphasizes the focus of such innovations to support medium and small businesses. The document also often mentions the so-called “recovery and resilience facility” in relation to artificial intelligence users. The term includes measures to support the EU population and business entities to more effectively master digital technologies, primarily artificial intelligence. The so-called “white paper” is expected to contain the most advanced and in-demand proposals in European society for the reform of legislation regulating relations in the field of artificial intelligence, the Internet of things and robotics. The second document, which is called the “Coordinated plan on artificial intelligence 2021 review” (hereinafter referred to as the Review), deals with more specific areas in which relevant activities are expected to be implemented. This document, which rightfully occupies a central place in the initiative we are considering, consists of four chapters: · Creation of favorable conditions for the application of artificial intelligence in the practical field; · Ensuring the efficiency of artificial intelligence production “from the laboratory to the market”; · Artificial intelligence as the key to improving living standards; · Artificial intelligence in the struggle for leadership in the field of mastering new digital innovations. The review includes a wide range of areas for more effective use of digital technologies, such as environmental protection and environmental management, public health, public administration, migration and protection of refugee rights. It seems that some areas, such as, for example, employment support and regulation of mobility of the able-bodied population, may be the subject of study by Russian specialists. And finally, the third document is the “Proposals for a regulation laying down harmonized rules on artificial intelligence” (hereinafter - Proposals). The key provision of this document is the idea of minimizing the risks that arise during the use of artificial intelligence. “The proposals are aimed at eliminating the risks associated with the specific use of artificial intelligence, classifying them into four different levels: unacceptable risk, high risk, limited risk and minimal risk,”8 the document’s preamble says. The significance of this document is quite comparable to the position of the Russian Concept discussed above. The document consists of two chapters. The first chapter outlines the proposals themselves, while the second contains numerous reference appendices. The proposals are drawn up in approximately the same format as Russian laws. In total, the Proposals contain 12 chapters and 85 articles, including numerous parts and paragraphs. In this paper, we will consider only, as we see it, the most significant chapters that may be of the greatest interest to a Russian researcher. The first chapter is devoted to the general provisions, it outlines the objectives of the Proposals and provides an interpretation of some legal categories. The second chapter includes areas of activity within which it is proposed to ban the use of artificial intelligence. For example, such an area is attempts to influence human consciousness with the help of artificial intelligence, as a result of which “involuntary changes in his behavior may be recorded”. It is also expected that the activities of various participants in legal relations, including public authorities, aimed at “assessing the reliability” of citizens, determining the scenario of their behavior in certain situations, will fall under the ban. In general, the provisions of this chapter seem to be aimed at protecting the interests and rights of citizens, as well as preventing the leakage of confidential information. The third chapter, which is called “High-risk artificial intelligence systems”, is perhaps the key in this document, since it considers those types of artificial intelligence that the European legislator classifies as a high-risk group. The appendix to the document provides a complete list of such digital technologies. For example, in the field of education, those innovations that affect “individuals’ access to educational institutions” may be prohibited, and in the field of migration policy, “artificial intelligence systems designed for use by competent government agencies as lie detectors, as well as tools for determining the emotional state of an individual”. The proposals envisage a special regime for monitoring the use of such types of artificial intelligence and tougher legal liability in case of violations of the law. Also, of particular interest to Russian researchers and legislators may be chapter five, called “Measures to support innovation”. This chapter focuses on the issue of supporting small and medium-sized businesses. In addition, the document clearly outlines those areas that will be considered priorities for public authorities to create the most favorable conditions for the use of artificial intelligence. The legislator attributed the prevention, detection and investigation of offenses, the field of healthcare and environmental safety to such areas. A detailed analysis of the main regulatory legal acts adopted in the EU has demonstrated that the key idea in the rule-making process is the desire of public authorities, on the one hand, to achieve maximum user trust, on the other hand, to ensure leadership in the field of introducing artificial intelligence into everyday human life. And our European partners are quite convincing when they state that “the European Commission is trying to find the necessary balance between supporting innovation by the population and ensuring that artificial intelligence benefits more than 500 million EU residents. If the prepared proposals are accepted, Europe may break away from the United States and China, which are not yet taking all the necessary steps to regulate the use of this digital technology” [2]. So, as can be seen from the above, the EU has not adopted a single and integral regulatory legal act regulating the use of artificial intelligence [3-10]. We reviewed the main regulatory acts and initiatives, namely: · Artificial Intelligence in Europe; · “The plan for coordinating activities in relation to artificial intelligence for 2021”; · “The European approach to artificial intelligence”. In our opinion, among these documents, the central place belongs to the “Proposals to the regulation establishing agreed rules on artificial intelligence”, which are an integral part of the European Approach we have considered. We believe that many provisions of the European Approach may be of interest to a Russian researcher. Comparative analysis A comparative analysis of the normative legal acts that were adopted in both legal systems at the time of this study allows us to draw certain conclusions, which will be outlined in the following order. At the beginning, those provisions will be outlined that indicate the similarity of approaches of Russian and European legislators, and then - demonstrating significant differences in the regulation of the use of artificial intelligence. The key provisions of the first group are the following: · At the time of writing, neither the Russian Federation nor the European Union had adopted a comprehensive, unified law regulating the use of artificial intelligence. This circumstance, in our opinion, is caused by the fact that the academic community of Russia and European countries has not yet developed an unambiguous understanding of the type of SCT we are considering. The ambiguity and lack of clarity in the interpretation of artificial intelligence is the main and extremely intractable task of modern jurisprudence. · Both in the Russian Federation and in the countries of the European Union, the legal framework within which it is supposed to formalize the regulation of the use of artificial intelligence is about the same. The initiator of this process is public authorities, not representatives of civil society. The fact is that the process of decentralization, which is the conceptual core of the new reality, is fraught with risks for the state structure to the greatest extent. As a result, the emerging digital law has a more developed administrative and legal mechanism for regulating such relations with its inherent imperative method of building legal relations. In our opinion, the most significant differences are as follows: · There is a clearer and more unambiguous understanding of end-to-end digital technologies, including artificial intelligence, in Russian legislation. For example, in the document “Artificial Intelligence in Europe” there are discrepancies in the interpretation of terms such as artificial intelligence, the Internet of things and robotics, their use as synonyms, or, conversely, their opposition. At the same time, as we see it, the understanding of artificial intelligence by the Russian authorities can be described as somewhat more progressive, appealing to the impossibility of human controllability of artificial intelligence. It seems that such an extremely progressive, or, in other words, ahead of time, approach can be revised to create a more comfortable atmosphere in the rule-making process for the Russian academic community and the legislator. · In the Russian legal field, the category of artificial intelligence sounds more clearly and consistently in the context of the progressive socio-economic development of the country [11; 12], while in the EU the emphasis is shifted in favor of protecting the rights of the user and defining the boundaries of his responsibility. · The activities of public authorities in the field of promoting the proper use of artificial intelligence are more thoughtful, structured and balanced. The EU’s legislative and law enforcement practice demonstrates the fact that government bodies lack a clear vision of the line that in the near future will determine not only the specifics of using digital achievements, but also the architecture of pan-European security in the virtual world. This circumstance is probably predetermined by the fact that European integration does not imply a centralized approach to solving such legal tasks. Another reason may be that the issues of goal setting, planning and forecasting have not been properly regulated in the European legal tradition. · In contrast to Russian legislation, the EU’s legislative process more clearly expresses the idea of appealing to so-called non-legal norms that take shape outside the framework of the main sources of law. A special place in this case is occupied by the norms of morality and the effect of public confidence in digital innovations. Conclusion At about the same time, the Russian Federation and the European Union began implementing the idea of the need to regulate the use of artificial intelligence by legal norms. The legislative process in both legal systems has a lot in common. For example, at the time of preparation of this article, neither Russia nor Europe had adopted a single regulatory legal act exclusively for this type of SCT. Moreover, a clear understanding of artificial intelligence still has not been formed. There are also the differences mentioned above. Apparently, the work of specialists and legislators in this area will continue in the format of approving strategies, approaches, plans and initiatives. It is likely that the institutionalization of artificial intelligence in the legal system of Russia and the EU is not a matter of the near future. As a kind of denominator for our research, I would like to say that the development of the world of digital transformations, as it turns out, is unpredictable in a certain sense. No matter how symbolic it may sound, artificial intelligence, as it turns out, is a kind of quintessence of this uncertainty. Moreover, the academic community and the legislator need to clearly and unambiguously understand that artificial intelligence is also a destructive technology with a very wide range of impacts. In the current situation, questions arise about how justified the intensive introduction of modern end-to-end digital technologies is, and how the dialogue between the social world and the technical world, between the state and civil society, will be further built. To answer such questions, there is a need to revise the concept of jurisprudence as a science and law itself as a purely social phenomenon. Apparently, the law is on the verge of its reformation, which will result in a shift towards a different understanding of the rule of law. It is possible that the very concept of the source of law and the normative legal act will be revised. Of course, in such a rapidly changing world, international experience should be of great interest to Russian researchers and legislators.

About the authors

Bulat D. Nuriev

State University of Management

Author for correspondence.
Email: nurievbd@mail.ru
ORCID iD: 0000-0001-8434-2419
SPIN-code: 5615-0500

Candidate of Philosophy, Associate Professor of the Department of Public and Municipal Management

99 Ryazansky prospect, Moscow, 109542, Russian Federation

References

Supplementary files