CC BY
42
UDC 349:681
BecTHMK CT6rY npaBO. 2024. T. 15. Bwn. 2
Artificial intelligence in law-making and law enforcement: Risks and new opportunities
D. A. Pashentsev1, Yu. G. Babaeva2
1 The Institute of Legislation and Comparative Law under the Government of the Russian Federation, 34, ul. Bol'shaya Cheremushkinskaya, Moscow, 117218, Russian Federation
2 Moscow Metropolitan Governance Yury Luzhkov University, 28, ul. Sretenka, Moscow, 107045, Russian Federation
For citation: Pashentsev, Dmitrii A., Yuliya G. Babaeva. 2024. "Artificial intelligence in law-making and law enforcement: Risks and new opportunities". Vestnik of Saint Petersburg University. Law 2: 516526. https://doi.org/10.21638/spbu14.2024.214
The article examines the use of artificial intelligence (AI) in law-making and law enforcement, analyzing it through the lens of the technological paradigm theory and utilizing a new methodology rooted in post-classical scientific rationality. The examination of the Al's legal personality takes into account the modern anthropocentrism standpoint, where humans are perceived as the central figure in the legal system. A human, as a subject of law, plays a crucial role in constructing legal reality. Through exercising his powers, a human transforms abstract norms into existing law. The proposal to delegate this anthropological function to AI carries significant risks that may have detrimental effects on the established system of legal regulation for public relations. The correlation between law-making and law enforcement, as integral components in establishing legal reality, pertains to the legal consciousness of individuals as subjects of law. This correlation does not necessitate the substitution of human decision-making with AI. Simultaneously, AI is capable of effectively handling auxiliary legal tasks. It possesses the necessary skills to draft regulatory acts, court decisions, and consolidate proposals obtained through law-making crowdsourcing. Additionally, it is proficient in executing automated actions, such as issuing writs of execution. The utilization of AI to implement the concept of machine-readable law and algorithmize the application of legal norms appears to hold promise. The incorporation of a digital state notion, which pertains to the transition of public service provision into digital format, is also unattainable without AI, which greatly enhances the public administration efficiency. In this case, the focal point will revolve around the matter of attaining effective control over decisions carried out by AI, as well as the prescribed boundaries of its capabilities within the legal domain, as established by legislative measures.
Keywords: artificial intelligence, law-making, law enforcement, subject of law, legal norm, digitalization, legal consciousness.
1. Introduction
In the modern world, the integration of new technologies is being actively pursued across all sectors of society and the state, prompting a transition to a novel technological paradigm. At present, the developed nations of the world have transitioned into the fifth technological paradigm. The shift to this paradigm was closely linked to the advent of computers and the proliferation of information technologies, as well as the emergence of
© St. Petersburg State University, 2024
the Internet, enabling humanity to maximize the simplification and acceleration of information exchange. The fifth technological paradigm has only been in existence for a few decades, yet indications of a new, sixth technological paradigm are already apparent. It is built upon sophisticated modern technologies that possess the ability to reshape all sectors of the economy and social landscape. To begin with, our focus will be on the widely recognized "big four", encompassing digital technologies, biotechnology, nanotechnology and cognitive technologies. Collectively, these technologies are commonly identified as convergent technologies, as their combined presence amplifies their impact on society. At present, digital technologies hold the top position in terms of development and are extensively utilized in public administration, public services, business and science, as well as in the daily lives of individuals within society. In light of this matter, along with the unique characteristics of the legal landscape, the advancements implemented within it during this current phase are typically linked to digital technologies.
The advent of digital technologies inherently brings up the question of employing artificial intelligence (AI). The recognition of the significance of this matter led to the adoption, in 2019, of the Decree of the President of the Russian Federation on the Development of Artificial Intelligence in the Russian Federation, which endorsed the National Strategy for the Development of Artificial Intelligence until 20301. On November 24, 2023, Vladimir Vladimirovich Putin, President of the Russian Federation, took part in the plenary meeting of the International Conference on Artificial Intelligence and Machine Learning "Artificial Intelligence Journey 2023" on the topic "The Generative AI Revolution: New Opportunities". In his speech, he expressed that the advancement of this technology heralds a new epoch in human history, rendering prohibitions in this domain irrelevant2.
The possibility of incorporating AI can also be observed within the legal domain. Legal scholars are currently engaged in active discussions regarding the extent and boundaries of such utilization in legal and law enforcement activities, as well as the trajectory of legal regulations, including the potential recognition of AI as a subject of law. Concurrently, the discussions draw upon previous studies pertaining to the role of digital technologies in the field of law, encompassing the potential risks and benefits associated with their implementation.
The examination of new societal processes influenced by technological advancements appears suitable for investigation using contemporary post-classical methodology. Post-classical rationality centers on the examination of intricate self-developing systems, while also adopting an interdisciplinary approach to comprehending the social and environmental aspects of reality. The complex methodological synthesis of AI is determined by its technological nature, which necessitates a dialectical understanding of its influence on social relations and the evolution of legal norms. The need to examine the impact of AI on law-making and law enforcement, from the perspective of the legal personality theory,
1 The Decree of the President of the Russian Federation No. 490 on the Development of Artificial Intelligence in the Russian Federation (together with the National Strategy for the Development of Artificial Intelligence until 2030) dated October 10, 2019. Sobranie zakonodatelstva RF. October 14, 2019. No. 41. Art. 5700. Available at "ConsultantPlus" system. Accessed May 30, 2024. https://www.consultant.ru. (In Russian)
2 The transcript of the speech of the President of the Russian Federation at the International Conference on Artificial Intelligence and Machine Learning "Artificial Intelligence Journey 2023" (November 24, 2023). The Official website of the President of the Russian Federation. Assessed December 8, 2023. http://www. kremlin.ru/events/president/transcripts/comminity_meetings/72811. (In Russian)
is determined by the synthesis of anthropocentricity in modern legal theory with constructivism as a post-modern approach to comprehending legal reality. The construction of legal reality occurs through three stages: law-making, the formation of legal consciousness and law enforcement (Pashentsev 2022). Legal consciousness, acting as a middle and thus crucial link in this process, is an inherent characteristic of a person. The unresolved inquiry into the potential existence of legal consciousness in AI resides beyond the realm of jurisprudence.
Thus, the present-day legal science is confronted with a crucial objective — to fully grasp the role that AI plays in shaping legal reality, in addition to its influence on law-making and law enforcement activities, and to ascertain the specific characteristics and limitations of its potential utilization within the legal sphere of society.
2. Basic research
In the context of transitioning to a new technological paradigm, the utilization of modern technologies is naturally expanding and being integrated into various spheres of society, state, and human life. Presently, there is a growing number of options for applying AI. These technologies encompass facial recognition, decision-making, transport management, and medical diagnostics. The integration of AI with Big Data technology has been instrumental, as it enabled the processing of vast quantities of information beyond human capacity. Every emerging opportunity, which materializes, necessitates the adaptation of legal regulations and the incorporation of new AI capabilities within the legal framework to anticipate and prevent potential conflicts and adverse outcomes. This has led to the growing significance of AI as a key force behind the evolution of law, influencing the trajectories of legal regulation. It is possible to anticipate that in the near future, it will not only aid individuals in performing legal, managerial, and other activities, but may also serve as a replacement, resulting in the displacement of various professions. This is also applicable in the context of law-making and law enforcement activities, where the utilization of AI is also growing.
The incorporation of AI into the legal framework is hindered by diverging interpretations of this phenomenon among legal professionals and technical specialists.
One of the legal definitions of AI is given in the National Strategy for the Development of Artificial Intelligence until 2030, which was approved by Decree of the President of the Russian Federation No. 490 dated October 10, 2019. As per this document, "artificial intelligence refers to a set of technological solutions that enable the replication of human cognitive functions (such as self-learning and problem-solving without a predetermined algorithm) and achieve results in performing specific tasks that are at least comparable to those of human intellectual activity. The set of technological solutions encompasses information and communication infrastructure, software (including those utilizing machine learning techniques), processes, and services for data processing and problem-solving". On the one hand, AI is undeniably defined as a set of technological solutions. On the other hand, perceiving AI solely as a ubiquitous technology employed across various domains of human endeavor is increasingly confronted with its unconventional capabilities, including self-learning and autonomous decision-making. This provides an opportunity for individual scientists to question the legal personality of AI, resulting in the emergence of a novel legal construct (Aksenova 2020).
Legal approaches to defining AI rely on isolating its essential characteristics, proceeding from the following indicators: the existence of a range of technological solutions; "the ability to reproduce human thought functions with comparable results; the ability to independently search for solutions and work without a predetermined algorithm; availability of software, including machine learning; organizational and technical implementation on the basis of information and communication infrastructure" (Guliaeva 2022, 60).
Various legal approaches exist regarding the definition of AI. As an illustration, an alternative method was utilized as the foundation for the initial European regulation on AI during the formulation of this concept. This approach entails the mere enumeration of technologies that meet the criteria of AI within the European Union and are employed in the creation of suitable software3. Consequently, the suggested definition appears excessively broad and ambiguous, rendering it challenging to implement in legal practice.
The U. S. Code also contains definition of AI: "artificial intelligence means any artificial system that performs tasks under varying and unpredictable circumstances without significant human oversight, or that can learn from experience and improve performance when exposed to data set"4. Such systems solve tasks requiring human-like perception, cognition, planning, learning, communication, or physical action (Begishev, Khisamova 2020, 708).
The doctrine has seen endeavors to establish and validate the concept of AI dating back to 1952. The definition formulated by American scientist John McCarthy encompasses the properties of robots, computer programs and systems to solve problems, draw conclusions, make decisions, and execute creative and intellectual human functions (Na-grodskaia 2019).
According to scientists, contemporary AI is referred to as "narrow" or "weak". "It consists of systems dedicated to specific tasks, based on models and algorithms designed and calibrated by human beings on the basis of data and parameters chosen at the discretion of the model-maker. This means that it is able to perform specific tasks, sometimes even better than human beings, but it cannot transfer those skills into other fields of knowledge. The term 'Intelligence' is, therefore, quite misleading" (Gallese Nobile 2023).
Simultaneously, efforts to define AI persist in the doctrinal realm, often surpassing legal definitions due to the doctrine's ability to adapt more flexibly amidst the rapid and accelerated advancements in modern digital technologies (Zaloilo 2021, 8). Nonetheless, it can be argued that firstly, there is a prevalent ambiguity in the various approaches towards comprehending artificial intelligence. This ambiguity not only impacts the perspectives of individual lawyers, but also gives rise to divergent viewpoints between legal professionals and those from technical and information fields. Secondly, it is important to note the existence of various definitions that encompass common characteristics. Primarily, AI is interpreted as a distinct technology, program, or set of programs capable of learning and performing cognitive tasks through extensive information processing. One crucial attribute of AI is its capacity for autonomous thinking and action, signifying a distinct level of independence and self-learning.
According to P. M. Morkhat, AI is characterized by the following essential qualities:
3 "Proposal for a Regulation laying down harmonised rules on artificial intelligence (April 24, 2021)". European Commission. Accessed May 30, 2024. https://ec.europa.eu/newsroom/dae/items/709090.
4 "Artificial Intelligence (10 U. S. Code § 2358)". Cornell Law School. Accessed May 30, 2024. https:// www.law.cornell.edu/wex/artificial_intelligence_(ai).
— virtuality that defines AI as a combination of software and hardware, encompassing information content along with the physical one;
— information and communication system can be defined as cybernetic, cyberphysi-cal, as well as bio-cybernetic;
— anthropomorphism of cognitive actions, self-regulation and the ability to structurally accumulate information. This includes the concept of genetic search, when information is stored in a ranked form and divided into different "generations" and "parental information" (Morkhat 2018, 5-8).
Additionally, M. V. Zaloilo holds the belief that "AI can be broadly defined as the capacity of intelligent systems to carry out creative and intellectual tasks that are inherent to humans" (Zaloilo 2021, 13).
The issue of incorporating artificial intelligence in the legal sector is not addressed, as its integration is already underway within the context of the overall digitization of law.
The term "digitalization of law" refers to all the processes occurring in the legal field due to the impact of digital technologies. It is a multifaceted phenomenon. The following areas of study within legal science can be considered in this context: the substantive impact of digital innovations on the evolution of law, leading to the emergence of new legal institutions; formal influence exerted by digital instruments on the procedural and procedural aspects of legal activities; potential impact of the digital environment on the legal value system (Dorskaia 2020). AI is just one component within a spectrum of digital technologies. However, when considering the magnitude of its impact on social relations and the legal landscape, particularly in the future, it is this technology that holds a prominent position, thus garnering the attention of legal scholars and practitioners.
In parallel, various new issues emerge from the ongoing processes, stimulating active discussions and leaving some of them quite debatable. Principal among these issues are those surrounding the potential legal personality of AI, particularly in relation to its role in judicial decision-making and the drafting of legal texts, such as regulatory acts.
Law-making is the initial phase in the establishment of a new legal reality. Consequently, deliberations concerning the implementation of AI commence specifically with regards to its potential utilization in the formulation and examination of draft regulatory acts.
According to Iu. G. Babaeva, the initial stage in introducing AI in lawmaking necessitates the development of technology of machine-readable norms (Babaeva 2022, 49). N. F. Poryvaeva correctly emphasizes the importance of distinguishing between the machine-readable format of regulatory acts and the law that is directly machine-readable (Poryvaeva 2021, 706). In the first case, the machine mechanically processes the text without delving into its meaning. In the second case, we are talking precisely about "clarifying" the meaning of legal norms. Therefore, machine-readable law encompasses legal norms that are formatted in a way that machines (computers) can interpret. This involves presenting the norms as algorithms written in programming languages and executed by machines without human intervention.
Translating the entire legal array into a machine-readable form seems to be a challenging task. Firstly, the challenge lies in the extent of current legal standards, as well as various linguistic, legal, technical, and mental issues (Pashentsev 2020). Secondly, the integration of AI in law-making and law enforcement pertains to the inadequate theoretical and legal examination of the relevant issues. Within this particular context, researchers
emphasize the necessity of addressing various challenges, such as enhancing the conceptual framework; determining the heterogeneity of legal statuses of various technologies of self-learning systems and robots equipped with AI; developing ways to distinguish responsibility for the consequences of AI actions; assessment of existing and potential risks, etc. (Guliaeva 2022, 59).
In relation to the process of legislation, we are currently discussing the utilization of AI as a supplementary technology to enhance the efficiency of drafting and implementing regulatory acts. In this context, several possibilities for such utilization can be pointed out. Firstly, this technology can be utilized for the preparation of draft regulatory acts. Today, robots are capable of successfully handling the writing of overviews, the preparation of analytical materials, the drafting of statements of claim, and other simple legal documents. Taking all of this into consideration, one can assert that AI has the ability to effectively solve the problem of preparing a draft regulatory act. Simultaneously, owing to its rapid data processing capabilities, it can formulate a draft in a manner that preemptively avoids potential conflicts with the current regulatory framework, thereby enhancing the quality of the proposed legislation. Secondly, AI can be employed to analyze legislation for the purpose of identifying current deficiencies in regulatory oversight, with the intention of addressing them in the future. Consequently, legal activity will adopt a more focused nature. Thirdly, the development of dedicated digital platforms has facilitated the implementation of a kind of law-making crowdsourcing, enabling extensive public discourse on the proposed regulatory legal acts. If a substantial number of diverse proposals are received, AI can be employed to effectively process these results. This technology can summarize the proposals to emphasize the main directions of public initiatives and formulate concise amendments. Ultimately, this approach aims to enhance the quality and effectiveness of law-making.
The application of AI offers expansive possibilities for the integration of legal forecasting as a significant tool in law-making policy.
Legal forecasting entails seeking a preliminary assessment, grounded in scientific evidence, of the potential outcomes resulting from the implementation of a specific regulatory act or norm. By means of forecasting, one can predict the progression of social relations and, consequently, undertake anticipatory law-making. In circumstances where rapid societal changes occur due to digitalization, legislation struggles to keep up and, therefore, becomes an inadequate means of regulation. Forecasting plays a crucial role in mitigating the impact of negative trends and enhancing the relevance of the adopted regulatory acts. Nevertheless, its potential remains largely untapped.
The assurance of forecasting efficiency lies in the consideration of its multifactorial nature, encompassing a wide range of conditions and circumstances that impact social development. Thus, AI, with its ability to process vast amounts of data, can be of immense value in this matter. It takes into account the data when constructing a forecast, thereby enhancing its objectivity and reliability.
The efficacy of law-making can be heightened by employing both forecasting and modeling. Considering the perspective of legal practice, a comprehensive digital model that encompasses the future development of the legal system, while addressing current problems and risks, will be of significance to lawyers. Employing it in legal activities will assist in the more comprehensive assessment of priorities while endeavoring to enhance the existing legislation, ultimately allowing for the refinement of its regulatory potential amidst the mounting global instability.
The notion of a digital law unveils vast prospects for incorporating AI in law-making. Such a law will not be drafted in a paper form or as a single approved text. It will be published online, which enables participants in public relations to flexibly apply and modify its norms based on changing life realities and their agreed-upon needs. A digital law will turn from a formally limited text into a convenient method of delivering relevant legal information to the recipients, presenting various alternative scenarios for them to choose from. Through the implementation of AI, it will be feasible to establish, monitor, and authorize the modifications of such a law, while mitigating potential risks, such as unauthorized interference with an electronic text.
The remarkable potential of AI in law-making suggests even greater possibilities for its utilization in law enforcement.
When discussing the digitalization of law enforcement, the primary focus is on judicial activities. In this context, two aspects are considered. The first one entails the application of innovative technologies in the preparation and conduct of a court hearing. Such efforts are being undertaken in foreign nations such as Brazil, China, the United States, and others. They illustrate that robots powered by AI successfully handle the preparation of court documents and the summarization of presented evidence.
Regarding this matter, it is worth noting the initial implementation of digitalization in judicial activities in Moscow, where there have been initial endeavors to incorporate technological advancements in court proceedings. These include the automatic recording in courtrooms (albeit with some challenges) and the conduct of online court hearings. The near future holds promise for the more active involvement of robots in tasks such as document and material preparation, as well as in the issuance of court orders. Consequently, these innovations have the potential to bring about substantial transformation in the judicial system, alleviate the workload of judges, and enhance the efficiency of the courts in Moscow. The auxiliary role of AI is evident in this context, without altering the general principles and procedures for administering justice. The main goal is to accelerate the processes in the judicial system by automating tasks that do not necessitate decisions on an alternative basis.
The second direction pertains to the projected application of AI in the process of judicial decision-making. Such attempts are also being made worldwide. Nonetheless, the general consensus among researchers and practitioners is that AI cannot entirely replace a judge. This is primarily because a judge's decision-making process encompasses not only legal norms and existing judicial practice, but also fundamental legal principles such as justice and integrity.
The consideration of employing AI for judicial decisions is linked to the advantages it offers. The following factors are included: the robot, in contrast to a human, does not get tired, does not require time for meals and rest, and is capable of quickly processing court cases 24/7. Given the widespread problem of overload in most courts, this solution seems to be an appropriate response. Moreover, AI remains unaffected by external and internal factors, devoid of any negative mood, incapable of being influenced by bribery or coerced into making unlawful decisions.
Compliance with the letter of law enhances the perceived legitimacy of the AI's justice. Apart from the letter, the law also encompasses the spirit that is manifested through the principles of law, such as justice, integrity, etc. The judge, in making a decision, relies on a personal assessment of the circumstances, his/her understanding of justice, finally,
he/she may have sympathy or antipathy for the parties to the process, based on a subjective assessment of their personal qualities, behavior and appearance. From one perspective, this could be perceived as exceeding the limits of the legitimacy, while from another perspective, it constitutes an indispensable element of justice, taking into account the entire historical progression of humanity within the realm of its anthropological nature. The absence of a human judge can lead to the erosion of morality in the realm of justice, as well as the diminishing reliance on the "people's spirit" that forms the fundamental nature of law, and is transmitted through a social relay. Court decisions devoid of an emotional component will become overly formalized, straightforward, insufficiently flexible.
In light of this matter, it is suggested to contemplate a more lenient approach towards incorporating AI into the justice system. This approach would ensure that the authority to make final law enforcement decisions, particularly in cases that require subjective judgment and assessment, remains solely with a human. AI has the potential to serve as a knowledgeable and intellectual assistant. It can analyze vast amounts of legal information, detect evolving patterns in judicial practices, locate relevant decisions and precedents, evaluate legislative norms, and predict the potential outcomes of specific decisions (Morkhat 2019, 68). Consequently, with the aid of AI, court decisions will be rendered more justified and efficient, resulting in a slight alleviation of the workload on judges.
The consideration of AI as a judge for complex cases in contemporary circumstances is primarily seen in relation to the future. The current level of AI does not instill enough confidence to use it in this capacity. "Such functions can only be assigned to complex AI systems, capable of performing complex tasks, but so far, perhaps, not existing, not achievable for the current level of development of science and technology in the field of AI, since simple mechanisms such as, for example, pattern matching-based, unable to model judicial decision-making, as the court must make reasonable and acceptable decisions often in those cases, when facts and norms, and how they interact with each other, are contradictory" (Morkhat 2019, 69).
The potential application of AI in law enforcement decision-making raises two interrelated problematic aspects: its legal personality and liability.
In contemporary legal science, legal personality is perceived as an innate quality of a person, derived from his/her personality and its legal status, comprising legal capacity, legal competence, and delictual capacity. It determines the position of the person in the legal space and in specific legal relations, implies its potential to be a subject of liability.
The concept of legal personality is intricately connected with the notion of legal consciousness, comprising legal ideology and legal psychology. A person who does not have legal consciousness cannot be recognized as sane. AI possesses a certain level of legal knowledge, yet lacks the capacity to experience emotions in connection with the law, thereby rendering it incapable of potentially embodying legal consciousness. Therefore, the inquiry into the legal capacity of AI cannot be adequately addressed in practical terms without compromising established legal structures, the very conception of personality in law, and its obligation to abide by the law.
The question of the AI's legal personality is a complex matter, leading to the existence of multiple perspectives expressed in doctrine. Their main ones are the following:
— AI is recognized as an object of law, a thing, thus rendering it subject to the existing norms of civil and information legislation, including laws on intellectual property;
— AI is just as much a fiction as it is a legal entity (Begishev, Khisamova 2020, 711);
— AI falls within the same legal framework as a person, and therefore, similar legal regulations should be applied to it;
— AI represents a new category of subjects of law, necessitating the development of specialized legal regulations (Zaloilo 2021, 40).
In 2017, a draft federal law on amendments to the Civil Code of the Russian Federation in terms of improving the legal regulation of relations in the field of robotics was developed and presented for discussion5. There was a proposal to add a chapter to the Civil Code that focuses on robots acting as agents intended for participation in civil circulation. It has been observed that such robots have the ability to possess the set-apart property, assume liability for their obligations, obtain and exercise civil rights on their own behalf, and assume civil obligations, while also participating in the civil proceedings. Therefore, the aim was to confer civil legal capacity upon the robot acting as an agent, enabling it to participate in civil relations similar to the norms on legal entities or property (whenever the robot assumes the role of an object of a transaction or is involved in relations concerning the liability of the owner of the source of increased danger for harm caused by the latter). The rights, obligations and responsibilities of the owner and actual keeper of the robot acting as an agent were determined. Concurrently, a robot was understood as "a device capable of acting, determining its actions and assessing their consequences based on information coming from the external environment, without full control by a human". It has been determined that the norms suggested for incorporation into civil legislation are not applicable to computer programs that are not integrated into the information system of a distinct device designed to execute actual actions autonomously, either fully or partially (Zaloilo 2021, 42).
The matter concerning the liability of AI has shifted from being solely theoretical to becoming a practical concern. This can be attributed specifically to the advancement of unmanned transportation systems. There may be occasions where, as a result of an erroneously decision made by AI, harm to property, health, or even life may occur, thus necessitating the establishment of legal regulations to address liability concerns. Under current legislation, a person can be held liable only if he/she has delictual capacity, that is, not only has he/she reached a certain age, but he/she can also be aware of the consequences of his/her actions, that is, he/she is sane. In order to bring charges against a person, it is imperative to establish his/her guilt through either intent or negligence. In criminal law, guilt is defined as the mental attitude of a person to the act committed by him/her. As AI lacks a psyche, it is unable to possess any mental attitude towards its decisions. In case of doing damage to property, the guilt is implied in the very fact of doing it, but AI, not possessing legal personality, cannot own property, therefore, cannot compensate for the damage done.
The draft law sparked debate within the scientific community and was ultimately rejected. Notable among the interesting proposals expressed during the discussion is the viewpoint of Judge G. A. Gadzhiev from the Constitutional Court of the Russian Federation. He suggests that if AI achieves a level of development where it exhibits signs
5 "The draft federal law on amendments to the Civil Code of the Russian Federation in terms of improving the legal regulation of relations in the field of robotics (2017)". Research Center for the Problems of Regulation of Robotics and Artificial Intelligence. Accessed May 30, 2024. http://robopravo.ru/uploads/s/ z/6/g/z6gj0wkwhv1o/file/My74kFFZ.pdf. (In Russian)
of consciousness and will in their legal sense, it should be deemed a quasi-subject of law (Gadzhiev 2018).
Consequently, the issue of the liability of AI itself cannot be raised in a practical realm. Henceforth, our focus remains solely on the liability of a person who assumes the position of owner, developer or operator of this AI. However, in this case, if we are talking about the criminal or administrative liability of this person, it becomes imperative to establish his/her guilt, which may not be possible if the decision was solely made by AI.
Therefore, the legislative decision pertaining to the potential legal personality of AI, albeit partial, appears premature. For this, the necessary theoretical, legal and practical-oriented prerequisites have not been formed. The hasty attribution of legal personality to AI has the potential to dismantle the existing framework of legal regulation of public relations and undermine the societal legal system. In turn, without solving this issue, which is also related to the issues of liability, it is also not possible to talk about the full independent participation of AI in law enforcement decision-making.
The study indicates that the use of AI in law-making and law enforcement appears to be objectively advantageous, although subject to certain guidelines and prerequisites. Such guidelines can be based on the following requirements:
— the use of AI solely within the strict confines of the existing regulatory framework;
— a combination of legal and ethical regulators;
— a primary focus on enhancing the level of protection and fulfillment of human rights and freedoms, while ensuring utmost attention to the well-being of a person;
— a scientific expertise of any decisions made in this field, along with an assessment of their consequences and risks;
— the definition of boundaries and subjects of liability in relations pertaining to the use of AI;
— the protection of personal data, including biometric ones;
— the legislative definition of the object and boundaries of the scope of AI application.
Hence, the incorporation of AI within the legal field is not an isolated pursuit but
rather an endeavor to enhance law enforcement and law-making processes, improve the quality of adopted regulatory acts, and increase the efficiency of law enforcement decisions, all with the intention of further rationalizing public affairs based on legal foundations and strengthening the legality and the rule of law.
3. Conclusions
The incorporation of AI in the legal landscape, along with its expanding utilization in law-making and law enforcement, is evidently a prevailing trajectory influenced by the transition to a novel technological paradigm and the integration of advanced technologies throughout society and the state. Concurrently, before embarking on the use of AI to address a specific legal matter, it is crucial to assess the practicability of such employment. Certain procedures within the realms of law-making and law enforcement occur in the absence of AI and, by their very nature, do not require an overly technological approach. Furthermore, in certain instances, AI has the potential to decrease rather than enhance the efficacy of legal activities. In cases where the use of AI is justified and logical, it becomes imperative to accurately define the boundaries of its application and establish efficient oversight by humans to ensure that crucial decisions impacting the well-being of individuals and societal
progress are made responsibly. To illustrate, AI will play a significant role in the future by controlling the formulation of draft regulatory acts in law-making. When it comes to law enforcement, it is imperative to exercise caution in employing this technology and refrain from solely relying on it to make specific court decisions. This is particularly important in cases where accurate assessment of both the case's circumstances and the individuals involved is necessary, along with adherence to principles of law. The key factor in ensuring maximum benefit and minimal risk of AI in the legal field lies in establishing a clear distinction between human and AI functions. At the same time, the legislative determination regarding the legal personality of AI seems, at the very least, premature.
References
Aksenova, Marina A. 2020. "The electronic entity concept in the legal framework". Iurist 7: 18-24. (In Russian) Babaeva, Iuliia G. 2022. "Digitalization and robotization in the focus of law". Vestnik Universiteta Pravitel'stva
Moskvy 1: 47-52. (In Russian) Begishev, Il'dar R., Zul'fiia I. Khisamova. 2020. "Artificial intelligence and robotics: Theoretical and legal problems of differentiation of the conceptual apparatus". Vestnik Udmurtskogo universiteta. Ekonomika ipravo 30 (5): 706-713. (In Russian) Dorskaia, Aleksandra A. 2020. "Transformation of law in the context of digitalization of public relations: Crisis phenomena and new opportunities". Transformatsiia i tsifrovizatsiia pravovogo regulirovaniia v sovremennykh realiiakh i usloviiakh pandemii. Ed. by Irina V. Vorontsova, 11-15. Kazan, Otechestvo Publ. (In Russian)
Gadzhiev, Gadis A. 2018. "Whether the robot-agent is a person? (Search of legal forms for the regulation of
digital economy)". Zhurnal rossiiskogo prava 1: 15-30. (In Russian) Gallese Nobile, Chiara. 2023. "Regulating smart robots and artificial intelligence in the European Union".
Journal of Digital Technologies and Law 1: 33-61. https://doi.org/10.21202/jdtl.2023.2 Guliaeva, Polina S. 2022. "Quasi-legal personality of artificial intelligence: Theoretical and legal aspects". Vestnik Moskovskogo gorodskogo pedagogicheskogo universiteta. Seriia: Iuridicheskie nauki 2: 58-69. (In Russian)
Morkhat, Petr M. 2018. Legal personality of artificial intelligence units. Civil Law Study. Moscow, IuNITI-DANA Publ. (In Russian)
Morkhat, Petr M. 2019. "The use of artificial intelligence in the judicial proceedings". Vestnik grazhdanskogo
protsessa 9 (3): 61-85. (In Russian) Nagrodskaia, Viktoriia B. 2019. New technologies (blockchain/artificial intelligence) in the service of law.
Moscow, Prospekt Publ. (In Russian) Pashentsev, Dmitrii A. 2020. "Features of law enforcement in the conditions of digitalization of social
relations". Vestnik of Saint Petersburg University. Law 11 (1): 35-49. (In Russian) Pashentsev, Dmitrii A. 2022. "Constructivism in modern legal science". Vestnik Moskovskogo gorodskogo
pedagogicheskogo universiteta. Seriia: Iuridicheskie nauki 1: 16-22. (In Russian) Poryvaeva, Natal'ia F. 2021. "Prerequisites for the emergence of machine-readable law". Iuridicheskaia
tekhnika 15: 706-709. (In Russian) Zaloilo, Maksim V. 2021. Artificial intelligence in law: A scientific and practical guide. Moscow, Infotropik Media Publ. (In Russian)
Received: December 12, 2023 Accepted: January 19, 2024
Authors' information:
Dmitrii A. Pashentsev — Dr. Sci. in Law, Professor; [email protected] Yuliya G. Babaeva — PhD in Law; [email protected]
