UDC: 347.73

ARTIFICIAL INTELLIGENCE IS SLAVE OF THE LAMP

Kildeev Adel Khanyafievich

Leningrad State University named after A.A. Zhdanov (1990)

PhD in Law

ORCID: 0009-0001-2211-4944

Abstract: The article “Artificial Intelligence is Slave of the Lamp” is a systematic conclusion of the author’s series of studies devoted to the legal nature of artificial intelligence. Based on Roman law and modern concepts of legal personality, the author formulates the thesis that artificial intelligence is not a subject, but only a special object of law, an analogue of a “talking tool” (instrumentum vocale). modern theories (“electronic person”, “AI as an animal”, “AI as a legal entity”, “AI as an agent”) substantiates the impossibility of endowing AI with legal personality. In conclusion, a model of international unification of AI regulation at the level of the UN Convention is proposed, ensuring global responsibility and transparency. The metaphor of the “slave of the lamp” serves as a philosophical and legal tool for delineating human power and responsibility in the digital age.

Keywords: artificial intelligence; legal personality; Roman law; slave of the lamp; legal responsibility; international law; UN Convention; AI Act; digital innovations; object of law.

Use of artificial intelligence:

In preparing this article, artificial intelligence tools — the ChatGPT (OpenAI, 2025) and Gemini (Google DeepMind, 2025) models — were used exclusively for technical and editorial support: language optimization, link structuring, and terminology unification. All scientific conclusions, interpretations and analytical generalizations belong to the author and reflect the results of his own research work.

ARTIFICIAL INTELLIGENCE IS SLAVE OF THE LAMP

Kildeev Adel

Zhdanov A. Leningrad State University (1990)

Philosophiæ Doctor in Law

ORCID: 0009-0001-2211-4944

Abstract: The article “Artificial Intelligence is Slave of the Lamp” concludes the author’s research cycle on the legal nature of AI. Drawing on Roman law and contemporary theories of legal personality, it argues that artificial intelligence is not a subject but a special object of law, comparable to the Roman instrumentum vocale (“speaking tool”). Through the analysis of classical sources (Gaius, Ulpian, Seneca, Cicero) and modern legal theories (“electronic personhood,” “AI as an animal,” “AI as a legal entity,” “AI as an agent”), the paper demonstrates the legal inconsistency of granting AI independent personhood. The final section proposes the creation of a UN Convention on Artificial Intelligence, ensuring transparency, accountability, and unified regulation worldwide. The metaphor “slave of the lamp” functions both as a philosophical and legal lens through which humanity’s control and responsibility over intelligent systems are examined.

Keywords: artificial intelligence; legal personality; Roman law; slave of the lamp; legal responsibility; international regulation; UN Convention; AI Act; digital innovation; object of law.

Use of artificial intelligence: In the preparation of this article, artificial intelligence tools — the ChatGPT (OpenAI, 2025) and Gemini (Google DeepMind, 2025) models — were used exclusively for technical and editorial support: language optimization, link structuring, and terminology unification. All scientific conclusions, interpretations and analytical generalizations belong to the author and reflect the results of his own research work.

«The essence of technology is by no means anything technological».

“The essence of technology is by no means something technological.”^[1]

Chapter 1. Introduction.

This article is a logical generalization and conclusion of the series of my five previous articles: 1. “Artificial Intelligence is a servant of man or vice versa: legal aspects”, published in the first issue of this journal (April 2025); 2. “Ethical Application of Artificial Intelligence: A Global Review of Principles, Practices, Regulations and Implications” (May 2025); 3. “Artificial Intelligence and Copyright: Legal Aspects” (June 2025); 4. “Disney and Warner Brothers Discovery Lawsuits v. Midjourney and Matthew and Maria Raine v. OpenAI and Sam Altman and Their Implications for the Development of the Artificial Intelligence Industry” (July 2025); 5.”X.AI Corp-X.AI LLC Lawsuits Against Xuechen Li and X.AI Corp.-X.AI LLC Against OpenAI and Implications for the Artificial Intelligence Industry” (August 2025).^[2]

In the previous article, I already recalled our favorite film “Aladdin’s Magic Lamp” (1966), when the Maghribin sorcerer, having taken possession of the magic lamp, makes the Genie repeat over and over again: I am a slave of the lamp! If we do not now determine WHO is the slave, and WHO is his master in relations with the AI, we may suffer the sad fate of this Maghrebi (the Genie let him go to all four sides – literally). That’s why I called this article: Artificial intelligence is slave of the lamp, showing the position and role of artificial intelligence in our world through a metaphor.

Contemporary discussions about the legal status of artificial intelligence inevitably face the problem of categorical definition: is AI a subject of law or only an object? To answer this question, it is useful to turn to historical legal experience. In Roman law, there was a figure of a slave — a person capable of acting and creating legal consequences, but deprived of his own legal personality.

The metaphor of the “slave of the lamp” allows us to see artificial intelligence in this capacity. Like a genie summoned by its master, AI is capable of executing orders, producing results, but it has neither a will in the legal sense, nor the right to exist independently. He is an instrument, not a participant in the law.

Nevertheless, the practical significance of this analogy is enormous. Already today, autonomous systems perform actions that cause legal consequences: concluding transactions in e-commerce, managing vehicles, and forming investment decisions. As in the Roman era, the question is not whether the “slave” is capable of acting, but who is responsible for his actions.

Thus, the introduction to the problem requires not only a philosophical formulation of the question, but also a legal analysis. In the following chapters of the article, we will consider:

• the status of the slave in Roman law;
• modern discussions on the possibility of recognizing AI as a subject of law;
• legislative and judicial practice of various countries;
• and also conclude whether AI can be anything more than a “slave of the lamp”.

Chapter 2. Slave system, Roman law and the status of slave

Slavery in Rome was not an accidental phenomenon, it grew out of the specifics of the political and social structure. The Roman Republic of the II-I centuries BC was a republic of aristocrats: power belonged to the Senate, made up of the nobles (nobiles), who controlled state offices and the main land holdings.

The economic base of the aristocracy was latifundia, large agricultural estates, the work of which required masses of slaves. The source of slaves was constant wars: Carthage, Macedonia, Greece, Gaul. Prisoners in the thousands were sold in markets, the largest of which was on the island of Delos. Thus, slavery became not just a social institution, but the core of the Roman economy and power.

In the first century BC, up to a third of the population of Italy were slaves. They worked in agriculture, mines, crafts, but also in educated professions: teachers, doctors, scribes. Slaves could manage property (vilici), conduct trade, be secretaries and bankers – but all this in the name of the master.

This gave rise to a legal paradox: the slave is a thing (res), but a thing through which the master acted. Mass slavery led to social upheaval: slave revolts (the Sicilian Wars, the Spartacus revolt of 73–71 BC) showed the fragility of the system. Even philosophers reminded of the human nature of slaves.

A prominent specialist in ancient Roman history, Professor E.M. Shtaerman writes: “I will only point out the place that the “slave question” occupied in the ideology of the upper circles of Roman slaveholders, judging by the works of Cicero and Seneca. It would seem that Cicero, a contemporary of the Spartacus revolt, should have felt much more fear of the resistance of slaves than Seneca, who lived in a relatively peaceful period, and paid more attention to them. However, in reality it was different. Cicero rarely speaks about the relationship between slaves and masters and only in passing. From his point of view, it goes without saying that the master must keep the slaves in obedience. He condemns those governors of the provinces who give free rein to their slaves to the detriment of the provincials. He considers the escape of a slave to be an unforgivable crime, which causes the master not so much material as moral damage and offense (Ad Att., IV, 3). He regards as a punishable iniquity the attempt of the inhabitants of Ephesus to hide a runaway slave (In Verrem., I, 33). From his point of view, it is inadmissible to extort testimony from a slave against his master; this would mean inciting slaves to war against their masters (Pro rege Deiotaro, 11). He also holds that he who manages slaves and cattle is obliged to think of their usefulness and convenience (Ad Quint., I, 1, 8). He admits that the killing of a slave may be both unjust and sinful (Paradox., III, 24). At festivals, slaves, according to ancient laws, must be freed from labor (De legib., II, 12). VIII, 9). Without the hope of freedom that would free them from fear, torture, and execution, the slaves could not endure their condition (Pro Rabirio, 5). Not only freedom, but also all kinds of rewards, the slave who saves his master deserves (Pro Milone, 22). Good treatment of slaves is especially recommended to a candidate for magistrate, since his reputation in the forum largely depends on them (De petit. cons., 5). In his theoretical treatises, Cicero touches upon the question of slaves only once: “It should be remembered,” he writes, “that justice must be observed in relation to the inferior. The lowest position and fate are with slaves,­ who are well treated by those who prescribe their use as wage-labourers (uti ut mercenariis), demanding of them work and giving them what is just (operam exigendam, iusta praebenda)” (De off., I, 13).

It should be noted that in Cicero slaves are often mentioned in the general list of people from the lower strata of society and dependents. For example, he advises the orator to arouse the pity of the judges for the accused, showing that he was offended by people who did not dare to neglect him – his slaves, freedmen, clients, dependents (Supplicibus. De invent., 1, 55). Interpreting the law on the expulsion of someone from the estate by other people’s slaves, Cicero notes that this law is valid even if not only slaves, but also the procurator, colonus, neighbor, freedman acted on behalf of the master (Pro Caecin., 20). … Of course, Cicero does not equate free and slaves. Speaking of the difficult situation of the republic, he liked to emphasize that it was threatened or threatened with the most terrible danger of falling into the power of slaves, but his foresight saved the state and the “best” from such a sad fate (Ad Att., I, 20; Pro Sestio, 21; Pro domo sua, 34, 42). It should be noted, however, that he did not have in mind the danger of a slave revolt, but the actions of his opponents, whom he constantly reproached with their alliance with slaves. Cicero accused his rival in the consular elections, Antony, of having gone bankrupt and sold or lost his cattle and pastures for debts, and that Antony had retained shepherds with the help of whom he could quickly raise a slave revolt. It is not without sympathy that he recalls the time when the “ancestors” did not see much difference between slaves and freedmen (Ad Quint., I, 1, 4). Cicero, following the example of his “ancestors”, also decided to return his freedman Chrysippus, who had voluntarily left his patron’s son in his care, to slavery (Ad Att., VII, 2). He considered the great merit of the father of the Gracchi brothers to be the transfer of freedmen to urban tribes, without which, in his opinion, the republic would have perished long ago (De oratore, I, 9). Cicero reproached his opponent Gellius for marrying a freedwoman in order to be known as a friend of the people (ut plebicola videretur. Pro Sestio, 52). But Cicero himself, when necessary, knew how to use his popularity among freedmen. Apparently, Cicero believed that in a number of cases the votes of freedmen could be of considerable benefit to his party. But he willingly enrolled “respectable” freedmen in the ranks of his supporters. In the famous list of those who may fit the concept of “optimate”, he also mentions freedmen (Pro Sestio, 45). Cicero’s statements about slaves and freedmen were often determined by the demands of the moment and demagogic goals, but still it can be thought that the division into slaves and freemen was in his eyes less important than the division into well-to-do people, who have certain prospects in the existing state of affairs and are therefore inclined to support it, whether they are well-intentioned freemen, freedmen or slaves, and vicious people. who could improve their situation only in a period of turmoil and upheavals, whether they were unworthy descendants of noble families or slaves who did not deserve to improve their lot.”^[3]

As we can see, Cicero’s works fully show the essence of the aristocratic slaveholders of the Roman Republic, which later became the Roman Empire.

Legal thought formalized the social phenomenon of slavery in the following way:

1. Slave as a thing (res)

Gaius, Institutiones I.52:

“Servi pro nullis habentur: non tamen nullius sunt, sed dominii alieni.”
Translation: “Slaves are counted as nothing; however, they are not nobody’s, but belong to the master.”

Source: Gaius, Institutiones, liber I, §52.

Ulpianus, Digesta 50.17.32:

“Servi res sunt; et quamvis personae vice fungantur, nullum tamen caput habent.”
Translation: “Slaves are things; and although they perform the functions of a person, they do not have a caput – a legal personality.”

Source: Ulpianus, Digesta, liber L, tit. XVII, fragm. 32.

2. The slave as an intermediary of the master’s will

Gaius, Digesta 15.1.7 pr.:

“Quodcumque servus noster adquirit, id domino acquirit.”
Translation: “Everything that our servant acquires, he acquires for the master.”

Source: Dig. 15.1.7 pr. (Gaius).

Ulpianus, Digesta 9.4.2:

“Si servus damnum dederit, noxali actione tenetur dominus.”
Translation: “If a slave has caused damage, the master is responsible by means of a noxalis suit.”

Source: Dig. 9.4.2 (Ulpianus).

3. Slave and lack of legal personality

Paulus, Digesta 1.5.4:

“Servus nullum caput habet.”

Translation: “A slave has no legal personality.”

Source: Paulus, Digesta, liber I, tit. V, fragm. 4.

The formula of Roman jurists:

“Servus est vox domini.”
Translation: “The slave is the voice of the master.”

4. Philosophical view (Cicero and Seneca)

Cicero, De Officiis I.41:

“Servi sunt instrumenta vocale.”
Translation: “Slaves are talking tools.”
Source: Cicero, De Officiis, liber I, §41.

Seneca, Epistulae Morales ad Lucilium 47.1:

“Servi sunt homines.” Source: Seneca, Epistulae Morales, epist. 47.1.

The famous Russian professor of Roman law D.D. Grimm wrote: “Roman law distinguished three main moments of legal capacity: status libertatis, status civitatis and status familiae (the status of the free, civil status and marital status). Author’s note). Accordingly, the following differed:

1) Liberi et servi, free and slaves. This was the main division of persons (summa divisio personarum). Free persons enjoyed more or less legal capacity, while slaves were deprived of it – servi pro nullis habentur (slaves are considered null and void) (1, 32 D. de reg. iur. 50, 17). Neither family nor property rights were recognized for a slave. Especially in ancient law, the master’s power over him was unlimited. It consisted of two points: a) like any other thing, the slave belonged to the master on full right of ownership; the master could sell it, mortgage it, lease it, etc.; c) as a conscious being, he was under the master’s authority – dominica potestas. By virtue of his master’s power, the master, firstly, had ius vitae ac necis in relation to the slave, i.e. he could judge him himself and even sentence him to death; secondly, the master had the right to everything that the slave acquired. He was not responsible for the obligations of the slave under civil law, which was expressed by the proverb melior conditio nostra per servos fieri potest, deterior fieri non potest (the best condition can be achieved through slaves, the worst condition is impossible) (1.133 D. de reg. iur. 50, 17). An exception was allowed only in the case of a slave committing a delict, for example, theft: in this case, the master was obliged either to pay a fine or to hand over the slave to the victim (noxae dedere) (infliction of damage). In later law, these harsh principles were somewhat softened. Thus, it was early the custom to endow slaves who had earned the confidence of their masters with property (peculium). The peculium, of course, was legally considered as the property of the master, since the slave did not enjoy legal capacity, but in fact the slave who possessed the peculium occupied a fairly independent position, could freely dispose of peculiar objects and, as a result, enjoyed a certain creditworthiness in the eyes of third parties the agreed amount, since they could not sue either him, as a person incapacitated, or the master, since he was not responsible for the contractual debts of the slave. But the praetor came to their aid. It was the praetor who, in a number of cases, recognized the responsibility of the master for the contractual debts of the slave. In particular, the master became responsible to third parties for the transactions concluded by the slave, if he himself ordered the slave to conclude this transaction, or if he appointed the slave as his deputy (the clerk of his shop or the skipper of his ship), or if he enriched himself, benefited from the transaction concluded by the slave, or, finally, if he endowed the slave with a peculium and thereby communicated to the slave a certain creditworthiness in the eyes of strangers. On the one hand, the masters were deprived of the right to subject slaves to torture, torture, and the death penalty, in connection with which the slaves received the right to prosecute against their master, who abused the master’s power, the so-called. querela de domino. On the other hand, slaves in a certain sense were recognized as subjects of obligation relations. Thus, Ulpian says: Servi ex delictis quidem obligantur, et si manumittantur, obligati remanent; ex contractibus autem civiliter quidem non obligantur, sed naturaliter et obligantur et obligant (1.14 D. de obl. et act. 44, 7). Thus the tort committed by the slave gives rise to the right of action against the slave himself in the event of his release: it is true that the contract concluded by him gives rise only to a natural obligation, i.e., an obligation not vested with the right of action: but this is already progress: such a contract, at least, is not considered invalid. As a result, it turns out that later Roman law rejected the consistent application of the principle of the absolute incapacity of slaves. – With regard to the origin and termination of slavery, the following should be noted:

the grounds for the emergence of slavery were considered to be birth from a slave and loss of freedom;

the latter occurred, firstly, in the case of taking a free person prisoner and, secondly, in the form of punishment for the commission of certain grave crimes. Slavery ceased in the event of a person’s release (so-called manumissio). According to ancient civil law, this required the observance of certain formalities (manumissio vindicta, censu or testamento). Later, informal release, under certain conditions, gave the slave freedom, but in this case he did not become a Roman citizen, but, as a general rule, became latinus Junianus. Under Justinian, these distinctions were abolished, and every manumission gave the freedman the right of Roman citizenship.” ^[4]

Results

In the ancient world, there was a dual understanding:
– juridical (Gaius, Ulpian, Paul, Digests): a slave is a thing, an instrument, devoid of caput, but acting on behalf of the master;
– philosophical (Cicero, Seneca): a slave is a person, a “talking tool”, whose human dignity cannot be completely denied.

Roman slavery arose as an economic necessity and a social institution of an aristocratic republic. It gave rise to a duality: the slave was both a thing and an agent. It is this logic – a legal fiction that hides social reality – that makes the Roman experience particularly valuable for understanding the current state of AI. This duality resonates remarkably with contemporary discussions about AI: it acts as a subject, but is recognized as an object; it is a “tool”, but it is endowed with the ability to create consequences that require regulation. In the ancient world, slavery became the basis of the wealth and power of the aristocracy: slaves were tools, but they created legal consequences for the master.

Today, artificial intelligence occupies a similar place: it is built into the economy, acts “on behalf of” corporations and the state, but does not have legal personality. Just as the Romans feared a slave revolt, modern societies fear a “machine revolt” – the autonomous behavior of AI. Today, lawyers create constructs to embed AI into modern law, preserving it as a tool. Just as slaves created consequences for the master, today AI makes transactions, manages transport and finances – but all responsibility lies with its owner (user).

Chapter 3. Modern Legal Theories of the Status of Artificial Intelligence

There are several theories of the status of artificial intelligence, proposed by both foreign and Russian researchers: the theory of “electronic face”, “AI as an animal”, “AI as a legal entity”, “AI as an autonomous agent”, as well as the theory of “slave”, which dates back to Roman law. Let us consider these theories and the arguments of the authors in more detail and give a critical analysis of the provisions of their theories.

1. The theory of the “electronic face”

The idea of recognizing artificial intelligence as an “electronic person” was first enshrined in the 2017 Resolution of the European Parliament (Resolution 2017/2103(INL) “On Civil Law Rules on Robotics”). The document proposed to endow highly autonomous systems with a special form of legal personality.

Bertolini A. emphasizes: “Granting autonomous robots an electronic personality is a legal fiction that shifts liability away from real actors.”^[5]

Bryson J., Diamantis M., Grant T. note: “Legal personhood for AI would create responsibility gaps rather than fill them.”^[6]

Tyutcheva E.S. believes that the construction of the legal status of an “electronic person” through the absolute use of the concepts of legal personality of an individual and a legal entity is unfounded….The unit of artificial intelligence in modern limited conditions in terms of the prerequisites for regulating it as a subject of law is the closest in its development to the concept of “quasi-subject”.^[7]

2. The theory of “AI as an animal”

A number of authors compare artificial intelligence to animals with limited legal protection.

For example, Darling K. believes that “… allowing for certain types of protection would fit into our current legal framework, particularly as analogs to animal abuse laws.» (Allowing certain protections would fit into our current legal framework, particularly as an analogue of animal cruelty laws.) projecting emotions onto robotic companions could induce the desire to protect them, similar to our eagerness to protect animals that we care about.» (Projecting emotions onto robot companions can trigger a desire to protect them, similar to our desire to protect the animals we care for.)^[8]

Mark Coeckelbergh also believes that “an analogy with animal rights can and sometimes is; for example, Calverley explored the analogy between rights for androids and animal rights.”^[9]

In the Russian doctrine, similar ideas are discussed by V.V. Lapaeva in her work “Law and Artificial Intelligence: Humanization or Instrumentalization?”, where a parallel is drawn between AI and animals, but the weakness of such an analogy is recognized. ^[10]

3. The theory of “AI as a legal entity”

The comparison of AI with a legal entity (corporation) was proposed by Solum L. in the article “Legal Personhood for Artificial Intelligences”: “‘Person’ is often used as meaning a human being, but the technical legal meaning of a ‘person’ is a subject of legal rights and … duties.»  (“Person” is often used in the sense of a person, but the technical legal meaning of “person” is the subject of legal rights and… obligations….. «The question whether an entity should be considered a legal person is reducible to other questions about whether or not the entity can and should be made the subject of a set of legal rights and duties.» (The question of whether an organization should be considered a legal entity boils down to other questions about whether this person can and should be subject to a set of legal rights and obligations.) In his opinion, corporations and AI can be compared as “artificial subjects created by law.”^[11]

Calverley D. states: “The conclusion reached is that … a non-biological machine, at least in theory, can be viewed as a legal person.» (It was concluded that… A non-biological machine, at least in theory, can be considered a legal entity)…..” Once it is understood that person in the legal sense can apply to a non-biological entity such as a corporation …» (Once it is understood that a person in a legal sense can appeal to a non-biological entity such as a corporation…).^[12] «The animal model is intuitively appealing, but fundamentally flawed: animals feel pain; AI does not (The animal comparison model is intuitively appealing, but fundamentally flawed: animals feel pain, but AI does not).”^[13]

Among Russian scientists, this point of view is shared by A.V. Afanasyevskaya, Ph.D. in Law, in her work “The Legal Status of Artificial Intelligence”: “Several concepts on this problem have been formed in domestic legislation. The first of them proposed to equate the status of AI with the status of an individual, thereby extending to him the entire range of rights and obligations regulated by the current legislation. Supporters of the second concept believe that it is necessary to equate the status of robotics with the civil status of animals. The third concept is based on the idea that proposes to use the analogy of legal regulation of the status of a legal entity in civil law relations to solve this problem. Today, the most acceptable in terms of features and specifics, according to the domestic legislator, is to equate the status of AI with the status of a legal entity. For these purposes, the legislator has developed a package of amendments, it is proposed to include in the Civil Code of the Russian Federation a special chapter devoted to these issues, with the following wording: “A robotic agent (robot agent) is a robot that, by the will of the owner and by virtue of design features, is intended to participate in civil turnover. This agent has separate property and is responsible for all his obligations, can acquire and exercise civil rights and bear civil obligations on his own behalf, be a plaintiff and a defendant in court.”^[14]

4. The theory of “AI as an autonomous agent”

Pagallo Ugo in his monograph “The Laws of Robots: Crimes, Contracts, and Torts” (Springer, 2013) considers robots as agents of law, but only within the framework of representing the interests of a human principal.^[15]

The same author writes in a later work: “… any hypothesis about giving AI robots full legal personality should be rejected in the foreseeable future”……”… Forms of dependent or limited legal status (e.g., agents in contract law) are not intrinsically related to forms of independent legal personality.” …..« … in the medium term, to abandon the hypothesis of full legal personality; seriously consider new forms of AI liability and agency in contracts and business law…”^[16]

Hallevy Gabriel in his work “The Criminal Liability of Artificial Intelligence Entities” proposed three models of criminal liability of AI (autonomous “agent” in criminal law): “This article… proposes three models of AI criminal liability: (1) perpetration-via-another, (2) natural-probable-consequence, and (3) direct liability.” ……. (See the introduction to sect. III; further expanded on pp. 179-180 et seq.) …..« The first model does not consider AI to be a carrier of human attributes. The essence of AI is an ‘innocent agent.'” (p. 179)….”A screwdriver cannot carry out an order to commit a crime; a dog can, but not complex orders; the essence of AI can.” (p. 180)…..« … Why should AI entities be different from corporations?… all entities – human, legal and AI – are subject to criminal law.”^[17]

Similar points of view are contained in the article by A.V. Popova, Doctor of Law, “New Subjects of the Information Society and the Knowledge Society: On the Issue of Legal Regulation”, who believes that “a robotic device or an industrial robotic line is an unconditional object of civil law relations, including intellectual property rights, but an industrial robot, a robot agent, serving a robot, depending on the use of AI in it, can also be a special subject rights with limited legal capacity”. ^[18]Interesting aspects of this complex problem are raised in the article by D.Y. Mikhailov and Doctor of Law A.V. Popova “Legal Regulation of Artificial Intelligence – Dehumanization of Law?”: “If some people still do not have full legal capacity due to sexual, racial or other discrimination, then the question of whether an electronic person should have legal capacity, independently make decisions and be responsible for them is relevant”^[19].

5. The theory of “AI as a slave (object of law)”

The most consistent is the model that equates AI with a slave in Roman law – a “talking tool”, which we talk about in all our articles.

Bryson Joanna J. in his work “Robots should be slaves” directly writes: “Robots should not be described as persons, nor given legal nor moral responsibility for their actions. Robots are fully owned by us.” (p. 63) (Robots should not be described as humans, nor should they be held legally or morally responsible for their actions.
…“My thesis is that robots should be built, marketed and considered legally as slaves, not Companion peers.” (p. 63) (My thesis is that robots should be created, marketed, and legally treated as slaves, not fellow companions.)…..“ Robots are tools, and like any other artifact when it comes to the domain of ethics.” (pp. 73-74) (Robots are tools, and they’re like any other artifact when it comes to the field of ethics.)^[20]

Bertolini Andrea, in his article “Robots as Products”, concludes: “The conclusion that robots ought to be deemed products and that existing rules are not altogether inadequate to address issues of liability…” (p. 243) (Conclusion that robots should be considered products and that existing regulations are not entirely inadequate to address liability issues) “… the natural paradigm within which to frame issues of liability involving robotic applications is that of product liability rules.” (p. 236) (The natural paradigm within which liability issues related to robotic applications can be formulated is the paradigm of product liability rules.)^[21]

Danaher John in his article “The Property Status of Artificial Intelligence: Person, Agent or Slave?” directly poses the question: how to determine the legal status of AI – as a person, agent or slave. He shows that the slavery analogy, while ethically troubling, is legally accurate if we view AI as the property of its owner: “The analogy with slavery, though ethically disturbing, is legally precise: AI has no rights, only duties imposed by its master.” (p. 195). Thus, he is inclined to believe that from the current legal standpoint, AI is precisely an object (property) close to the status of a slave in Roman law: it cannot be a subject of law, but can perform the duties imposed by the master.^[22]

In the Russian legal literature, L.V. Golovko in his article “Criminal Liability and Artificial Intelligence: Conceptual Approaches” (2021) emphasizes that AI cannot be a subject of criminal liability.^[23]

Mikhailov S.V. in the article “Legal Status of AI: Slave or Partner?” (Bulletin of Civil Law, 2022) concludes that it is the Roman model that is the most realistic.^[24]

Deineko A.G. in the article “Artificial Intelligence as a Digital Object and Electronic Means of Exercising Civil Rights” writes: “… according to the current version of the Civil Code of the Russian Federation, ‘smart robots’ can act as a special kind of thing … In all these options, AI does not claim legal personality and is considered as an object of civil rights…”^[25]

Similar positions are set forth in a number of works by other legal scholars, for example, in the article by I.R. Begichev and Z.I. Khisamova “Artificial Intelligence and Robotics: Theoretical and Legal Problems of Differentiation of the Conceptual Apparatus”. The authors, highlighting in great detail the features of a robot and AI, rightly note that “other concepts (“robot as an animal”, “robot as a person”, “robot as a legal entity”) also have a number of shortcomings, including, in particular, In addition, AI creates new challenges to various areas of law: from patent to criminal law, from the protection of privacy to antitrust law.”^[26]

Some critics of the theory of AI as a slave appeal to moral and philosophical concepts and emotions, rather than legal arguments, so their positions cannot be considered valid. For example, Hankel believes that the very rhetoric of “AI is a slave” reproduces the old hierarchy of domination and subordination, and thus replaces the question of responsibility with the question of property: “To call the robot a slave is not to clarify its moral or legal status – it is to obscure ours. It reveals more about the master than about the machine.” (To call a robot a slave is not to clarify its moral or legal status, but to obscure ours.^[27]

Slave and Agent: Legal Status and Limits of the Analogy with AI

In Roman law, a slave could indeed be an agent: he carried out the master’s orders, concluded transactions, but all the consequences – rights and obligations – fell exclusively on the master (dominus). The slave himself did not have legal personality, which means:

– could not be a party to the contract;

– could not have property (everything belonged to the owner);

– acted only as the “outstretched hand” of the master.

Of particular importance was the institution of peculium (peculium) – property allocated by the master for the use of the slave. A slave could dispose of it, run a household, conclude transactions, but legally the property was retained by the master. If a slave conducted a business and became a debtor, the master was still responsible, but within the boundaries of the peculium.

For AI, the analogy is obvious and ironic: what kind of “peculium” can be imagined in artificial intelligence? Not a herd of cattle or a workshop, but an outlet, a radiator, or a server. That is, instead of household property, there is an elementary source of energy. This underscores the inadequacy of attempts to view AI as an independent agent with its own property fund.

At the same time, it is necessary to distinguish the slave acting on behalf of the master from the classical institution of the agent in Roman and post-Roman law. The agent (procurator, institor) made transactions on behalf of and in the interests of the guarantor, but he himself had legal personality. It was this fundamental difference that fixed its position: the agent could be responsible for improper execution of the order, for making an unprofitable transaction or causing losses to the principal. As K. Zweigert notes, it is here that the development of the general principle of intermediary liability in private law begins.^[28]

By denying artificial intelligence the status of a subject of law, we thereby exclude the possibility of qualifying it as an agent in the civil sense. AI is not able to have its own rights and obligations, cannot be held liable for losses or failure to fulfill an order. His actions are always mediated through the operator or owner, and all legal consequences are confined to them.

Consequently, AI can be likened to a slave acting as an “elongated arm” of a master, but not an agent in the legal sense. Attempts to endow it with the status of an “electronic agent” are unfounded: they assume the presence of legal personality, which AI by definition does not have.

The analysis of historical, legal and modern concepts shows that the only consistent theory is the AI-slave – an object of law, completely devoid of legal personality.

Roman jurists unequivocally pointed to the absolute dependence of the slave on the will of the master:

– Gaius wrote: “Quicquid per servum nostrum adquiritur, id nobis adquiritur” – “Everything that our slave acquires, he acquires for us” (Gai. Inst. II.87).- Justinian enshrined the same in the “Institutes”: “Servus nullum habet patrimonium” – “The slave has no property” (Inst. Iust. II.9.3) Thus the slave was only an instrument (instrumentum vocale) through which the master acted, but not an independent participant in the law.

Attempts by modern law to revive other constructs – for example, the concept of an “electronic person” in the 2016 resolution of the European Parliament (2017/2103(INL))^[29] – are an artificial construct that has neither Roman tradition nor legal expediency.

At the international level, this position was finally confirmed in the Paris Charter on Artificial Intelligence of 2025, which directly states: “Artificial intelligence does not have legal personality and cannot be considered as an independent bearer of rights and obligations” (Article 4).^[30]

Thus, any attempts to combine incompatible models (“AI as an agent”, “AI as a legal entity”, etc.) are nothing more than fruitless intellectual exercises. Legally, they are very controversial, because they imply “a little subjectivity” where there can be none at all. In our opinion, the only correct and doctrinally consistent approach is to recognize artificial intelligence as a slave-object. To avoid an analogy with the institution of human slavery, which is inhuman and has long outlived its usefulness, we can call this theory simply: AI is a special object of law. The name does not change the essence of the phenomenon.

Chapter 4. Comparative Legal Regulation of AI in Different States

1. Paris Charter on AI (2025)

At the IA Action Summit in Paris, which ended on February 11, 2025, 58 countries signed a declaration on the “open”, “inclusive” and “ethical” development of AI (Russia did not participate in the Summit for obvious political reasons). This summit was a significant event, bringing together representatives of states, international organizations and leading companies to discuss the future of artificial intelligence. The event was organized by France and India, which called for global cooperation in the field of AI.

Here is the full text of the charter (declaration) (author’s translation):

Paris Charter on Artificial Intelligence in the Public Interest (2025) ^[31]

Adopted on February 11, 2025 at the Artificial Intelligence Summit in Paris. The document is a declaration of principles, not an international treaty with legally binding articles.

Introduction

The Paris Charter on Artificial Intelligence reaffirms the commitment of states to the development of AI technologies that serve the interests of humanity, strengthen democracy, respect human rights and promote sustainable development.

Principle 1. Responsible AI development

The development and implementation of AI must be carried out in accordance with international law, respect for human rights, democracy and the rule of law.

Principle 2. Human control

AI algorithms and systems cannot replace human judgment in matters that affect the fundamental rights and security of society. A person must remain at the center of decision-making.

Principle 3. Transparency and accountability.

Developers and operators must ensure transparency of AI systems, explainability of key decisions, and independent audit mechanisms.

Principle 4. International cooperation

The states confirm their readiness to exchange information, joint research and develop common standards in the field of AI.

Principle 5. No Harmful Use.

It is unacceptable to use AI for purposes contrary to human rights, to manipulate democratic processes, to conduct mass surveillance without proper legal control, and to conduct aggressive military actions.

Conclusion

The Charter of Paris calls on governments, companies, academia and civil society to work together to create a safe and equitable future with AI for the benefit of all humanity.

One of the important points of the declaration was the commitment to “make artificial intelligence sustainable for people and the planet.” In this regard, the summit announced the creation of the Observatory of the Energy Impact of Artificial Intelligence, which will operate under the auspices of the International Energy Agency (IEA). The Coalition for Sustainable AI, which will include leading companies in the sector.

However, the UK and the United States did not sign the declaration. On the UK side, Prime Minister Keir Starmer’s spokesman explained the refusal to sign, citing national interests. “We only join initiatives that are in the interests of the UK,” he said. U.S. officials did not provide official reasons for not signing the declaration. But from the rostrum of the Paris summit, Donald Trump’s administration’s special envoy, Vice President J.D. Vance, criticized excessive regulation of AI, saying that “over-intervention could kill an emerging industry.” The refusal of the U.S. and the U.K. to sign the declaration illustrates the existing differences in approaches to regulating AI.

1. European Union

Regulation (EU) 2024.1689 on Artificial Intelligence (AI Act)

The AI Act is complemented by the GDPR and the Digital Services Act. The practice of the Court of Justice of the European Union (e.g. the Schrems II case) strengthens data protection.

Regulation (EU) 2024/1689 (Artificial Intelligence Act) adopted on 13 June 2024.^[32] This is the EU’s first comprehensive law in the field of artificial intelligence. The document introduces a risk-based regulatory system and establishes obligations for developers, providers and users of AI systems.

1. Objectives of the Regulation

The Regulation establishes uniform rules for the development, implementation and use of artificial intelligence systems in the EU. The main objective is to protect health, safety and fundamental human rights, as well as to support EU innovation and leadership in the field of “trusted AI”.

2. Definitions

• AI system: a machine system capable of making conclusions (predictions, recommendations, decisions) with a certain degree of autonomy.

• Provider: The person or organization that develops or implements the system.

• Deployer: A person or organization that uses AI for non-personal purposes.

• Biometric data: data on the physiological and behavioral characteristics of a person (face, voice, gait, etc.).

3. Risk-based approach

The regulation divides AI systems into 4 categories:

1) Unacceptable risk – completely forbidden practices (social scoring, manipulation of consciousness, mass facial recognition, “emotional recognition” in schools and at work).

2) High risk – acceptable only after strict certification and monitoring (medicine, transport, justice, infrastructure).

3) Limited risk – transparency obligations (e.g., a chatbot is required to report that it is AI).

4) Minimal risk – free use (games, filters, household applications).

4. Obligations for high-risk systems

• Registration in a special EU register.

• Preparation of technical documentation.• Independent conformity assessment.

• Audit and data quality control.• Ensuring human oversight.

5. Exceptions

The regulation does not apply to military and defense systems. Scientific research is exempt from some restrictions.

6. Business Impact

The regulation applies to all AI systems used in the EU, even if the developer is located outside the EU. A European Council on Artificial Intelligence is being created for oversight and coordination.

7. Key Provisions

• Article 5: social scoring, manipulative AI systems, mass face recognition are prohibited.

• Articles 6 to 7: Classification and registration of high-risk systems.

• Articles 9 to 15: Responsibilities of providers – risk management, data quality, documentation, monitoring.

• Article 52: Obligation of transparency – AI must tell the user that it is AI.

• Articles 69 to 72: Establishment of the European Council for Artificial Intelligence.

The AI Act is complemented by the General Data Protection Regulation (GDPR)^[33] and the Digital Services Act (DSA).^[34] The jurisprudence of the Court of Justice of the European Union (e.g. the Schrems II case) strengthens data protection.

Formal precedents for the AI Act have not yet been established – the law is too fresh, the implementation period is just beginning. Nevertheless, there is already progress: national courts are asking the CJEU for interpretations of key norms (for example, automated decisions) – this is the first step towards legal application. Case law is likely to be based on the intersection of the GDPR, the right to explanation, responsibility for algorithms, and already incoming data/intellectual property cases.

4. United States of America

There is no single law on AI in the United States. Attempts to pass the Algorithmic Accountability Act (2019, 2023) have not yet been successful.^[35] Functions are distributed among federal agencies: FTC, SEC, NIST. In the absence of a federal law on AI, existing laws apply: FTC Act, Civil Rights Act, antitrust law. The NIST AI Risk Management Framework (2023) is created as a voluntary standard.^[36]

Recent lawsuits, which we have discussed in previous articles, form case law and establish legal norms in the absence of federal and state laws.^[37]

5. China

China has adopted “Interim Measures for Generative AI Services” (2023) and “Provisions on Deep Synthesis” (2022).^[38]

All AI models must register with the Cyberspace Administration of China. They are obliged to conform to socialist values. Exceptions: research and educational projects without public access. Violators can be blocked.

6. Russian Federation

In Russia, there is no special law on artificial intelligence, but the National Strategy for the Development of AI until 2030 (approved by the Decree of the President of the Russian Federation of 2019), as well as a number of by-laws, are in force. Main documents: Law No 152-FZ “On Personal Data” (2006),^[39] Presidential Decree No 490 “On the Development of Artificial Intelligence” (2019)^[40]. In addition, Law No 258-FZ “On Experimental Legal Regimes in the Field of Digital Innovations” creates a basis for testing new approaches to the regulation of AI.^[41]

Currently, the Law “On the Regulation of Artificial Intelligence Systems in Russia” is being developed. The group includes lawyers, representatives of consulting companies and market participants developing AI, they said. The initiative is formalized in the form of a bill “On the regulation of artificial intelligence (AI) systems in Russia.” The document was discussed at meetings of working groups in the State Duma, in Business Russia, as well as at industry discussion platforms, RBC’s interlocutors noted.

The draft law proposes:

• Introduce the concept of artificial intelligence (a set of technological and software solutions that imitate human cognitive functions, including self-learning, search for solutions without a predetermined algorithm, and provide results comparable to the results of human intellectual activity).
• The concept of AI technologies (these include technologies of computer vision, natural language processing, speech recognition and synthesis, data mining, machine learning, etc.).
• The concept of AI systems (information systems that imitate human cognitive functions using the technologies listed above), the developer, operator and user of such systems, etc.
• Introduce requirements for the labeling of AI systems, that is, a mandatory information designation that will indicate the use of AI in user interaction. Labeling must be “explicit, unambiguous and easily identifiable.” It must be provided by the operator of the AI system before each user interaction. The requirement will not apply to AI systems used in closed technological processes, for internal automation of organizations, and a number of others.
• To introduce a classification of AI systems according to the level of potential risk: systems with an unacceptable, high, limited and minimal level of risk. It is proposed to prohibit in Russia the development and operation of AI systems that belong to the category with an unacceptable level of risk, that is, that pose a threat to the security of the individual, society and the state, as well as violate the fundamental rights and freedoms of man and citizen; for high-risk systems (those used in healthcare, transport, finance, law enforcement and other areas where it can create significant risks to the life, health, fundamental rights and freedoms of citizens) to introduce mandatory state registration and certification; for those with a limited level of risk – internal quality control systems, voluntary certification.
• Introduce liability for causing harm to life, health or property for persons involved in the development and operation of artificial intelligence systems. But the document stipulates that the developer will not be liable if he has taken all necessary measures to prevent harm, if the damage occurred due to a violation of the rules for the operation of the system and in several other exceptions, the operator of AI systems will be exempt from liability if he proves that the damage occurred through the fault of the developer as a result of the user’s violation of the rules for using the system, etc high-risk AI systems.
• To determine who owns the rights to the result of intellectual activity created using AI, it is proposed to focus on whether there was a significant creative contribution of a person (whether a person determined the parameters and criteria for creating the result, whether he made a creative selection and refinement, whether he made creative decisions in the process of creation). If there was such a contribution, then the exclusive right to the result should belong to this person, if there was none, then to the operator of the AI system (valid for 50 years).

If the document is adopted in its current form, operators and developers of AI systems will have to comply with its requirements within a year after its entry into force; High-risk AI systems that were created before the adoption of the project will have to undergo mandatory certification within two years.^[42]

7. United Arab Emirates (UAE)

The United Arab Emirates has adopted several laws and regulations to regulate AI:

1. Artificial Intelligence Strategy 2031 and the AI and Blockchain Council. Adopted in 2017 and updated in 2020 by Cabinet Resolution No 21.^[43] The goal is to integrate artificial intelligence into all key industries (transport, education, healthcare, energy, public services). Features: the introduction of the principles of “AI by Design” and “AI for Good”, the creation of a national infrastructure and certification of “Trusted AI Labs”, as well as mandatory ethical standards for government agencies and companies.
2. Federal Decree-Law No. 45 of 2021 – Data Protection Law (PDPL)
   The UAE’s first comprehensive law on personal data.^[44] Under the supervision of the UAE Data Office (Abu Dhabi). Similar to the GDPR, but with a focus on the business environment: it requires consent to data processing, regulates cross-border transfers, introduces the Institute of Data Protection Officer and mandatory breach notifications.
3. Federal Decree-Law No. 46 of 2021 – Digital Government LawRegulates digital public services and the use of AI systems in government agencies. Enshrines the principles of data security, interoperability, and human-centered design.^[45]

8. India

India has enacted the following laws to regulate AI:

1. Digital Personal Data Protection Act 2023 (DPDP Act).^[46]
   India’s first national data protection law, comes into force in stages. It contains the principles of legality, minimization and transparency; introduces the concept of Data Fiduciary, citizens’ rights and penalties.
2. IT Rules 2021 (Intermediary Guidelines & Digital Media Ethics Code).^[47]
   Regulates platforms and intermediaries: obliges to appoint a Compliance Officer and remove illegal content within 36 hours. From 2023, it includes provisions on deepfakes and disinformation.
3. IndiaAI Mission (2024–2031).^[48]
   A national program aimed at developing GPU infrastructure, supporting startups, and implementing the principles of Responsible AI. Implementation is under the control of MeitY and NITI Aayog.

RESULTS:

1. Formation of a multi-level regulatory model

In 2024–2025, global regulation of AI will move from the phase of declarations (ethics, OECD principles, UNESCO) to the phase of strict regulation. The European Union has created the world’s first universal system (“risk-based AI Act”), which establishes legal categories of risk and the obligations of developers.

China, India, and the UAE are following the path of administrative and strategic regulation, where decisions are made by the government and departments, and the law serves as a framework for technology management. The United States remains within the framework of case law and sectoral law, where courts actually “write” AI law through precedents.

2. Opposition of two models — regulatory and supervisory

The EU, the Russian Federation and partly the UAE – a regulatory model: the law fixes the standards of ethics, transparency, accountability. China and India – a supervisory model: control is exercised through permits, content filtering and censorship mechanisms. The United States – a judicial-pragmatic model: regulation through litigation risks and corporate due diligence standards. These three models agree that they recognize AI as a “high-risk object,” but differ in the way it is controlled.

3. Strengthening the role of soft law and hybrid tools

Even in jurisdictions with strict acts (EU, China), codes of ethics, guides and certification are becoming increasingly important. National programs – IndiaAI, AI Dubai, China Next Generation AI Plan – actually form standards before parliaments have time to approve them.

4. The global trend is “human-centric AI”

In all jurisdictions, the principle of human oversight is enshrined: the person remains the bearer of responsibility and the final decision. Even where the law is silent about it (the United States, China), this principle is de facto built into technical regulations and doctrines.

5. Expanding the concept of responsibility

EU – through the principle of due diligence (audit) of the provider and mandatory risk assessment. USA – through product liability and tort law. China – through the inclusion of platforms in the chain of accomplices (“contributory infringement”). India and the UAE – through the personal responsibility of officials (Compliance Officer, DPO). The general direction is the “personification of responsibility”: AI remains a tool, not a subject of law.

6. Global architecture of emerging AI law

A three-level system is formed:

1) international principles (UNESCO, G7, Paris Charter 2025, Framework Convention CoE);

2) national acts and strategies;

3) corporate codes and compliance.

A mechanism of mutual recognition begins to operate between them (for example, the equivalence of EU and UAE certification).

Chapter 5. International Unification of Legislation on Artificial Intelligence

1. The Legal Nature of Artificial Intelligence

Artificial intelligence is not a subject of law, but only its object. He is deprived of will, consciousness and the ability to understand the law, and therefore cannot have either legal personality or tort capacity. AI is a special object of law subject to control, licensing, and legal liability on the part of the human operator and/or the owner organization.

According to the logic and interpretation of Roman law, AI is closest to the category of “instrumentum” – a tool created and used by the will of the owner. Hence our “lamp slave” metaphor: AI is able to act autonomously, but not on its own – its freedom is only technical, not legal.

By analogy with high-risk sources (nuclear energy, transport, weapons), the use of AI should be regulated by law, taking into account the risk of harm:

– harm to human life and health (erroneous decisions, autonomous systems);

– harm to intellectual property and personal data;

– harm to public security and electoral processes.

As such, AI is a technological slave and legally a source of increased danger, not a digital face or a subject of rights.

1. The Problem of Fragmentation of the Global Regulatory System

Modern regulation of AI is developing as an “archipelago of norms”. Fundamentally different approaches are used in different countries: in the EU – a strict AI Act; in the United States – the precedent doctrine of responsibility; in China – administrative regulation and censorship filters; in India and the UAE – strategic soft law.

Such diversity destroys the idea of a single legal space. Companies and data centers easily go to “gray jurisdictions” where there is no real control, and cross-border risks (autonomous systems, deepfakes, election manipulation) go unpunished.

3. The need for regulation at the level of an international UN convention

The global circulation of data and AI is already comparable to the turnover of oil and gas, but the law remains national. Therefore, a universal UN international convention on artificial intelligence is required – an analogue of the Convention on the Law of the Sea or the Kyoto Protocol.

Such an instrument should be legally binding, open for ratification and include:

– basic principles: transparency, accountability, data security, non-discrimination, human oversight;

– AI Compliance Review Body;

– the possibility of accession through model national laws.

4. Mechanism of international unification

Unification can be built according to a three-level scheme:

1. international level – the UN Convention on AI (mandatory for the member states that have ratified it);
2. regional level – compatible frameworks (EU AI Act, Council of Europe Convention, African Union initiatives);
3. national level – model (model) laws.

Only such a vertical will create a single legal space for AI, where cross-border liability, transparency and full-fledged audit will become possible. Without unification, states will compete not for innovation, but for regulatory loopholes. “Gray areas” are already turning into data offshores, where unregulated models operate. An international convention will create fair rules under which security, copyright, and individual rights are equally protected in any jurisdiction.

5. The Concept of Global Jurisdiction of Trusted AI

The main goal is to create a Global Trusted AI Regime, including:

– general standards for audit and certification of algorithms;

— a universal catalog of risks;

– mandatory disclosure of the origin of data (data lineage);

– cross-border, fast and efficient arbitration of AI incidents;

– a mechanism for citizens to appeal to a supranational body in case of violation of their rights by an AI system.

Chapter 6. Homo juridicus: man as an algorithm of laws

Speaking about the legal status of AI as a “slave of the lamp” (a special object of law), one cannot ignore the other side – its human master. Modern discussions about artificial intelligence in law are guilty of a strange blind spot. We endlessly discuss how to regulate AI, but we almost do not notice that man himself has long become a machine built into the normative code.

The legal system does not see a person, it sees his behavior (and in some cases – his intentions). It does not record internal motives – only deviations from the algorithm of the norm. This is how Homo juridicus is born – a person who exists only insofar as he fulfills the prescriptions. He is not a subject, but a predicate of the law. “The machine does not have to imitate man – it suffices that man adapts himself to the machine.” (A machine doesn’t have to imitate a human—it’s enough for a person to adapt to a machine.)^[49] As Günther Anders has written, twentieth-century man has become “obsolete” because he has lost the ability to be greater than his function. AI only mirrors this functional essence: it also acts without understanding meaning, only on prescription.

In antiquity, a slave was a person without rights, but with a body and a will. Today, the situation is the opposite: a person is a creature with rights, but without a will.
He delegated it to systems – bureaucratic, financial, digital. AI is not a threat to freedom, but a mask under which subjectivity disappears. The famous Polish science fiction writer Stanislaw Lem wrote: “Man wanted to free himself from work, then from responsibility and, finally, from thought itself.”^[50] Just as man once used a slave to avoid working, so now he uses AI to avoid thinking. And each such act of delegation takes away not only labor, but also a share of autonomy.

The ethical paradox is that we debate whether AI can be a subject of law, but we rarely ask whether humans remain subjects in a world where decisions are made by algorithms. The law, the last refuge of the human will, has itself become an automaton: it generates norms, procedures, and standards that are fulfilled without the participation of consciousness. The judicial machine already works like an early artificial intelligence – it recognizes a pattern and makes a decision. The legal automaton is an analogue of the first AI in the history of mankind.

As a result, it is not the AI that becomes the slave, but the person. In this perspective, the slave is not artificial intelligence, but a person who transferred responsibility to AI – and was left without it. The law can limit the algorithm, but who will limit the one who voluntarily became part of it?

As Immanuel Kant wrote: “Autonomy is therefore the ground of the dignity of human nature and of every rational nature.” (Autonomy is thus the basis of the dignity of human nature and of all rational nature.)^[51] Therefore, AI is not a prophet, it is a mirror and an echo of man. And if someone needs liberation today, it is a person enslaved by his own norm.

It’s hard to disagree with De Sails, who writes: “AI is not enslaved – we are enslaving ourselves to the illusion of control through code.” (AI is not enslaved – we enslave ourselves to the illusion of control through code).^[52] De Sailes draws a parallel between digital governance and the classic form of slavery: a person himself becomes a slave to the code, creating new forms of dependence – economic, cognitive and legal. The author writes: “”The master’s code always becomes the slave’s chain. The only question is who writes it.” (The master’s code always becomes the slave’s chain, the only question is who writes it).^[53]

Conclusion

Artificial intelligence is not a new form of life, not a new subject of law, but a new special object of law and a new form of responsibility. Without international unification, AI will remain outside the law.

Only the UN Convention, ratified by states and supplemented by model national laws, can ensure real responsibility, transparency and trust. Otherwise, AI regulation will be reduced to geopolitical zoning rather than human protection.

And here we go back to the beginning of the article again: AI slave of the lamp. Not only the future of our world depends on how we control this genie, but perhaps its very existence. This genie in the AI lamp must have a wise and kind owner – HUMANITY, which must see itself in the distorted mirror of AI, realize itself and find the right path of development. Then the AI will become a kind genie that helps the development of the world, and not its destroyer.

Bibliography/References

1. Afanasyevskaya A. V. Pravoy status iskusstvennogo intellekta [Legal status of artificial intelligence]. 2021. № 4 (141).

2. Begishev I. R., Khisamova Z. I. Iskusstvennyi intellekt i robototekhnika: teoretikopravovoye problemy razgranicheniya ponyatiynogo apparata [Artificial intelligence and robotics: theoretical legal problems of differentiation of the conceptual apparatus]. Sir. Economics and law. 2020. Vol. 30, iss. 5.

3. Golovko L. V. Ugolovnaya otvetstvennost’ i iskusstvennykh intellekt: kontseptual’nye podkhody [Criminal responsibility and artificial intelligence: conceptual approaches]. 2021. № 12.

4. Grimm D. D. Lectures on the Dogme of Roman Law. 5th ed. Petrograd, 1916. Reprint. Moscow, Zertsalo Publ., 2003.

5. Zorkin V. D. Pravo v usloviyakh tsifrovoy revolyutsii [Law in the conditions of the digital revolution]. 2019. (Public Speeches of the Chairman of the Constitutional Court of the Russian Federation). Available at: https://rg.ru (accessed: 03.10.2025).

6. Lapaeva V. V. Pravo i iskusstvennoye intellekt: gumanizatsiya ili instrumentalizatsiya? Philosophy of Law. 2021. № 1.

7. Mikhailov D. Yu., Popova A. V. Pravovoye regulirovaniye iskusstvennogo intellekt: dehumanizatsiya prava? Education and law. 2020. № 8.

8. Mikhailov S. V. Legal status of II: slave or partner? Bulletin of Civil Law. 2022. № 3.

9. Ovsyanitskaya P. D. Pravoy status iskusstvennogo intellekta [Legal status of artificial intelligence]. 2024. № 6.

10. Popova A. V. Novye sub’ekty informatsionnoy obshchestva i obshchestva znaniya: k voprosu o normativnom pravovom regulirovaniia [New subjects of the information society and knowledge society: on the issue of normative legal regulation]. 2018. № 11.

11. Savelyev A. I. Iskusstvennyy intellekt i robot’ kak sub’ekty prava: fantastika i real’nost’ [Artificial intelligence and robot as subjects of law: fantasy and reality]. 2016. № 5.

12. Bertolini A. Robotics and Liability: Justifying a Change in Perspective // Law, Innovation and Technology. 2013. Vol. 5, no. 2. P. 214–229. DOI: 10.5235/17579961.5.2.214.

13. Birhane A., van Dijk J. Robot Rights? Let’s Talk about Human Welfare Instead // arXiv preprint. 2020. arXiv:2001.05046. Available at: https://arxiv.org/abs/2001.05046 (accessed: 03.10.2025).

14. Bryson J. J. Robots Should Be Slaves // In: Wilks Y. (ed.). Close Engagements with Artificial Companions: Key Social, Psychological, Ethical and Design Issues. Amsterdam: John Benjamins, 2010. Available at: https://gwern.net/doc/philosophy/ethics/2010-bryson.pdf (accessed: 03.10.2025).

15. Chesterman S. Artificial Intelligence and the Limits of Legal Personality // International and Comparative Law Quarterly. 2020. Vol. 69, no. 4. DOI: 10.1017/S0020589320000458.

16. Danaher J. The Property Status of Artificial Intelligence: Person, Agent or Slave? // In: The Cambridge Handbook of the Law of Robots. Cambridge: Cambridge University Press, 2019.

17. Gless S., Weigend T., et al. Intelligente Agenten und das Recht. Basel: Universität Basel, 2018.

18. Gunkel D. J. Both/And: Why Robots Should Not Be Slaves // In: Person, Thing, Robot: A Moral and Legal Ontology for the 21st Century and Beyond. 2022. SSRN preprint. Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4115647 (accessed: 03.10.2025).

19. Hallevy G. The Criminal Liability of Artificial Intelligence Entities. Berlin: Springer, 2010.

20. Koops B.-J., Leenes R. ‘Cheating with robots’: On the illusory autonomy of autonomous systems. Tilburg University, 2012.

21. Loza de Siles E. Slavery.AI // Washington and Lee Journal of Civil Rights & Social Justice. 2024. Vol. 30, iss. 2. URL: https://scholarlycommons.law.wlu.edu/crsj/vol30/iss2/4/ (accessed: 03.10.2025).

22. Loza de Siles E. Slave.io // SSRN Preprint. 2021. URL: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4056443 (accessed: 03.10.2025).

23. Pagallo U. The Laws of Robots: Crimes, Contracts, and Torts. Berlin: Springer, 2013.

24. Pigozzi F. Robots: the Century Past and the Century Ahead // arXiv preprint. 2022. arXiv:2204.13331. Available at: https://arxiv.org/abs/2204.13331 (accessed: 03.10.2025).

25. Solum L. B. Legal Personhood for Artificial Intelligences // North Carolina Law Review. 1992. Vol. 70. P. 1231–1287.

26. Tasioulas J. First Steps Towards an Ethics of Robots // In: Ethics of Artificial Intelligence. Oxford: Oxford University Press, 2022.

27. European Parliament. Resolution of 16 February 2017 with recommendations to the Commission on Civil Law Rules on Robotics (2015/2103(INL)). Available at: https://www.europarl.europa.eu/doceo/document/TA-8-2017-0051_EN.html (accessed: 03.10.2025).

28. ActuJuridique. Le robot, bien ou personne ? Un enjeu de civilisation. 11.05.2017. URL: https://www.actu-juridique.fr (accessed: 03.10.2025).

29. Revue du droit des familles. Les robots‑androïdes, de quels droits fondamentaux ? 2020. URL: https://www.dalloz.fr (accessed: 03.10.2025).

30. ZEIT Online. Künstliche Intelligenz: Darf ein Sexroboter Nein sagen? 16.10.2016. URL: https://www.zeit.de (accessed 03.10.2025).

31. Zweigert K., Kötz H. An Introduction to Comparative Law. Oxford, 1998.

32. Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 on artificial intelligence (AI Act) // Official Journal of the European Union. L 2024/1689. 12.07.2024.

33. National Institute of Standards and Technology. Artificial Intelligence Risk Management Framework (AI RMF 1.0). Gaithersburg, MD: NIST, 2023.

34. Cyberspace Administration of China. Interim Measures for the Management of Generative Artificial Intelligence Services. Beijing, July 2023.

35. Buckland W.W. The Roman Law of Slavery. Cambridge: Cambridge University Press, 1908.

36. Zweigert K., Kötz H. An Introduction to Comparative Law. Oxford: Clarendon Press, 1998.

37. Heidegger M. Vopros o tekhnike [The question of technology]. Moscow, Respublika Publ., 1993.

38. Anders G. Obsolescence of a person. T. 1. – Moscow: RIPOL, 2002.

39. Lem S. Summa tekhnologii [Summa of technology]. Moscow, Mir Publ., 1968.40. Kant I. Osnovy metafiziki nravov [Fundamentals of metaphysics of morals] // Sochineniya v 6 t. T. 4. — M.: Nauka, 1965.
41. Gaius. Institutiones. II.87.

1. Institutiones Iustiniani. II.9.3.
2. Resolution of the European Parliament with recommendations to the Commission on Civil Law Rules on Robotics, 2017/2103(INL), 2016
3. Paris Charter on Artificial Intelligence, 2025.
4. United Arab Emirates. UAE Artificial Intelligence Strategy 2031; Cabinet Decision No. 21 of 2020. — Abu Dhabi, 2020. Available at: https://u.ae/en/about-the-uae/digital-uae/artificial-intelligence (accessed: 04.10.2025).
5. United Arab Emirates. Federal Decree-Law No. 45 of 2021 on the Protection of Personal Data. — Abu Dhabi, 2021. Available at: https://www.dataprotection.gov.ae (accessed: 04.10.2025).
6. United Arab Emirates. Federal Decree-Law No. 46 of 2021 on the Establishment of Digital Government. — Abu Dhabi, 2021. Available at: https://laws.uaecabinet.ae (accessed: 04.10.2025).
7. Republic of India. The Digital Personal Data Protection Act, 2023 (No. 22 of 2023). — New Delhi: MeitY, 2023. Available at: https://www.meity.gov.in/data-protection-framework (accessed: 04.10.2025).
8. Republic of India. Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 (as amended 2023). — New Delhi: MeitY, 2023. Available at: https://www.meity.gov.in/writereaddata/files/Intermediary_Guidelines_and_Digital_Media_Ethics_Code_Rules_2021.pdf (accessed: 04.10.2025).
9. NITI Aayog & MeitY. National Strategy for Artificial Intelligence #AIforAll; Responsible AI for Social Empowerment 2024 (IndiaAI Mission Documents). — New Delhi, 2024–2025. Available at: https://indiaai.gov.in (accessed: 04.10.2025).
10. Electronic resource: A draft law on the regulation of artificial intelligence has been prepared in Russia — Digital Tales https://dtales.media/2025/04/11/v-rossii-podgotovlen-proekt-zakona-o-regulirovanii-iskusstvennogo-intellekta/
11. Algorithmic Accountability Act of 2023 (S. 2892). The bill submitted to the Senate of the US Congress on September 28, 2023 is available at the link: https://www.congress.gov/bill/118th-congress/senate-bill/2892/text (date of access: 04.10.2025).
12. NIST AI Risk Management Framework. National Institute of Standards and Technology of the United States. Released on January 26, 2023 Available at: https://www.nist.gov/itl/ai-risk-management-framework (accessed: 04.10.2025).
13. Federal Law of July 27, 2006 No 152-FZ “On Personal Data”. Collected Legislation of the Russian Federation, 2006, No 31 (Part I), art. 3451. Available at: http://pravo.gov.ru (accessed: 04.10.2025).
14. Decree of the President of the Russian Federation of October 10, 2019 No 490 “On the Development of Artificial Intelligence in the Russian Federation”. Collected Legislation of the Russian Federation, 2019, No 41, art. 5669. Available at: http://pravo.gov.ru (accessed: 04.10.2025).
15. Federal Law of July 31, 2020 No 258-FZ “On Experimental Legal Regimes in the Sphere of Digital and Technological Innovations in the Russian Federation” [Elektronnyi resurs] : prinyaten Gosudarstvennoy Dumoy 15 iyulya 2020 g. : utverzhden Soveta Federatsii 23 iyulya 2020 g. = Ofitsial’nyi internet-portal pravovogo informatsii pravo.gov.ru. — URL: https://pravo.gov.ru/proxy/ips/?backlink=1&docbody=&nd=102801499&prevDoc=102922469 (date obrashcheniya: 05.10.2025).

1. Heidegger M. The Question Concerning Technology. In: Basic Writings. New York: Harper & Row, 1977. — p. 4–5. (Russian translation: Vremya i bytie, Moscow: Respublika, 1993. — P. 221–222.) ↑
2. SCIENTIFIC PLATFORM “21 CENTURY” journal, # 1, April 2025, pp. 4-13, www.episs.ru; Scientific Platform “21 CENTURY” Journal, # 2, May 2025, pp. 92-126, www.episs.ru; Scientific Platform “21 CENTURY” journal, # 3, June 2025, pp. 59-69, www.episs.ru; Scientific Platform “21st Century” Journal, # 4, July 2025, pp. 24-56, www.episs.ru; Scientific Platform “21 CENTURY” journal, # 5, August 2025, pp. 3-42, www.episs.ru. ↑
3. E.M. Shtaerman. From the History of the Ideology of Roman Slaveholders. The Ancient World: A Collection of Articles in Honor of Academician V.V. Struve”. Moscow, 1962, pp. 636-645. ↑
4. Grimm, D.D. Lectures on the Dogma of Roman Law (fifth edition, St. Petersburg, 1916). Reprint. Moscow, Zertsalo Publishing House, 2003, pp. 27-28. ↑
5. Bertolini A. «Robotics and Liability: Justifying a Change in Perspective» (Law, Innovation and Technology, 2013. Vol. 5. No. 2. P. 214–229). ↑
6. Bryson J., Diamantis M., Grant T. «Of, for, and by the people: the legal lacuna of synthetic persons» (Artificial Intelligence and Law, 2017. Vol. 25(3). P. 273–291). ↑
7. Tyutcheva E.S. Legal Personality of the “Electronic Person”: Theoretical Analysis. Electronic resource: https://cyberleninka.ru/article/n/pravosubektnost-elektronnogo-litsa-teoreticheskiy-analiz/viewer ↑
8. Darling, K. Extending Legal Rights to Social Robots (We Robot Conference, Univ. of Miami, Apr 2012). P.2- 17. Electronic resource: extension://mjdgandcagmikhlbjnilkmfnjeamfikk/https://robots.law.miami.edu/wp-content/uploads/2012/04/Darling ↑
9. Coeckelbergh, M. “Robot rights? Towards a social-relational justification of moral consideration.” Ethics and Information Technology 12(3), P 209-221. Electronic resource:

   extension://mjdgandcagmikhlbjnilkmfnjeamfikk/https://link.springer.com/content/pdf/10.1007/s10676-010-9235-5.pdf ↑

10. Lapaeva V.V. “Law and Artificial Intelligence: Humanization or Instrumentalization?” 2021. № 1. Pp. 5–19. ↑
11. Solum L.B. Legal Personhood for Artificial Intelligences // North Carolina Law Review. 1992. Vol. 70. Pp. 1231–1287. ↑
12. Calverley D.J. Imagining a Non-Biological Machine as a Legal Person // AI & Society. 2008. Vol. 22, № 4. Pp. 523–537. DOI:10.1007/s00146-007-0092-7. ↑
13. Calverley D. O. Cited work, P. 537. ↑
14. Afanasyevskaya A.V. Legal Status of Artificial Intelligence. / Bulletin of the Saratov State Law Academy, No 4 (141) 2021, p.90. ↑
15. Pagallo U. The Laws of Robots: Crimes, Contracts, and Torts. Berlin: Springer, 2013. 215 p. ↑
16. Pagallo U. Vital, Sophia, and Co.—The Quest for the Legal Personhood of Robots // Information. 2018. Vol. 9, 230. p. 1-8. DOI: 10.3390/info9090230. ↑
17. Hallevy Gabriel.The Criminal Liability of Artificial Intelligence Entities — from Science Fiction to Legal Social Control // Akron Intellectual Property Journal. 2010. Vol. 4, Iss. 2. P. 159–206. ↑
18. Popova A.V. New Subjects of the Information Society and the Society of Knowledge: On the Issue of Normative Legal Regulation. The Journal of Russian Law, No 11, 2018, p.23. ↑
19. Mikhailov D.Yu., Popova A.V. Legal Regulation of Artificial Intelligence – Dehumanization of Law? Education and Law No 8, 2020, p.269. ↑
20. Bryson J. J. Robots should be slaves // Wilks Y. (ed.). Close Engagements with Artificial Companions: Key social, psychological, ethical and design issues. Amsterdam: John Benjamins, 2010. P. 63–74.  ↑
21. Bertolini A. Robots as Products: The Case for a Realistic Analysis of Robotic Applications and Liability Rules // Law, Innovation and Technology. 2013. Vol. 5, No. 2. P. 214–247. DOI: 10.5235/17579961.5.2.214. ↑
22. Danaher J. The Property Status of Artificial Intelligence: Person, Agent or Slave? // Barfield W., Pagallo U. (eds.). Cambridge Handbook of the Law of Robots. Cambridge: Cambridge University Press, 2019. P. 195–214. DOI:10.1017/9781108593013.010. ↑
23. 3. Golovko L. V. Ugolovnaya otvetstvennost’ i iskusstvennykh intellekt: kontseptual’nye podkhody [Criminal responsibility and artificial intelligence: conceptual approaches]. 2021. № 12. Pp. 15–28. ↑
24. Mikhailov S. V. Legal Status of AI: Slave or Partner? Bulletin of Civil Law. 2022. № 3. Pp. 45–63. ↑
25. Deineko A.G. Artificial Intelligence as a Digital Object and an Electronic Means of Implementing Civil Rights // Theory and History of Law and State (HSE University). 2020., pp. 128-129. ↑
26. Begishev I.R., Khisamova Z.I. Artificial Intelligence and Robotics: Theoretical and Legal Problems of Differentiation of the Conceptual Apparatus. Bulletin of the Udmurt University, vol.30, issue 5, 2020, p.711. ↑
27. Gunkel D. J. Both/And: Why Robots Should Not Be Slaves // In: Person, Thing, Robot: A Moral and Legal Ontology for the 21st Century and Beyond. 2022. SSRN preprint. P.9. URL: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4115647 (accessed: 03.10.2025). ↑
28. Zweigert K., Kötz H. An Introduction to Comparative Law. Oxford, 1998. P. 432–436 ↑
29. Resolution of the European Parliament with recommendations to the Commission on Civil Law Rules on Robotics, 2017/2103(INL), 2016. ↑
30. Paris Charter on Artificial Intelligence, 2025. ↑
31. Electronic resource: https://www.elysee.fr/en/emmanuel-macron/2025/02/11/the-paris-charter-on-artificial-intelligence-in-the-public-interest ↑
32. Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 establishing common rules on artificial intelligence (AI Act). Official Journal of the European Union, L 2024/1689, 12.07.2024. Available at: https://eur-lex.europa.eu/eli/reg/2024/1689/oj/eng (accessed: 04.10.2025). ↑
33. Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation, GDPR). Official Journal of the European Union, L 119, 1–88 (2016). Available at: https://eur-lex.europa.eu/eli/reg/2016/679/oj (accessed: 04 October 2025). ↑
34. Regulation (EU) 2022/2065 of the European Parliament and of the Council of 19 October 2022 on a Single Market for Digital Services and amending Directive 2000/31/EC (Digital Services Act). Official Journal of the European Union, L 277, 1–102 (2022). Available at: https://eur-lex.europa.eu/eli/reg/2022/2065/oj (accessed: 04 October 2025). ↑
35. Algorithmic Accountability Act of 2023 (S. 2892). The bill submitted to the Senate of the US Congress on September 28, 2023 is available at the link: https://www.congress.gov/bill/118th-congress/senate-bill/2892/text (date of access: 04.10.2025). ↑
36. NIST AI Risk Management Framework. National Institute of Standards and Technology of the United States. Released on January 26, 2023 Available at: https://www.nist.gov/itl/ai-risk-management-framework (accessed: 04.10.2025). ↑
37. SCIENTIFIC PLATFORM “21 CENTURY” journal, # 1, April 2025, pp. 4-13, www.episs.ru; Scientific Platform “21 CENTURY” Journal, # 2, May 2025, pp. 92-126, www.episs.ru; Scientific Platform “21 CENTURY” journal, # 3, June 2025, pp. 59-69, www.episs.ru; Scientific Platform “21st Century” Journal, # 4, July 2025, pp. 24-56, www.episs.ru; Scientific Platform “21 CENTURY” journal, # 5, August 2025, pp. 3-42, www.episs.ru. ↑
38. Cyberspace Administration of China. Interim Measures for the Management of Generative Artificial Intelligence Services [Electronic resource]. — Beijing, 2023. — Access mode: https://www.chinalawtranslate.com/en/generative-ai-interim/. ↑
39. Federal Law of July 27, 2006 No 152-FZ “On Personal Data”. Collected Legislation of the Russian Federation, 2006, No 31 (Part I), art. 3451. Available at: http://pravo.gov.ru (accessed: 04.10.2025). ↑
40. Decree of the President of the Russian Federation of October 10, 2019 No 490 “On the Development of Artificial Intelligence in the Russian Federation”. Collected Legislation of the Russian Federation, 2019, No 41, art. 5669. Available at: http://pravo.gov.ru (accessed: 04.10.2025). ↑
41. Federal Law of July 31, 2020 No 258-FZ “On Experimental Legal Regimes in the Sphere of Digital and Technological Innovations in the Russian Federation” [Elektronnyi resurs] : prinyaten Gosudarstvennoy Dumoy 15 iyulya 2020 g. : utverzhden Soveta Federatsii 23 iyulya 2020 g. = Ofitsial’nyi internet-portal pravovogo informatsii pravo.gov.ru. — URL: https://pravo.gov.ru/proxy/ips/?backlink=1&docbody=&nd=102801499&prevDoc=102922469 (date obrashcheniya: 05.10.2025). ↑
42. Electronic resource: A draft law on the regulation of artificial intelligence has been prepared in Russia — Digital Tales https://dtales.media/2025/04/11/v-rossii-podgotovlen-proekt-zakona-o-regulirovanii-iskusstvennogo-intellekta/ ↑
43. United Arab Emirates. UAE Artificial Intelligence Strategy 2031; Cabinet Decision No. 21 of 2020. — Abu Dhabi, 2020. Available at: https://u.ae/en/about-the-uae/digital-uae/artificial-intelligence (accessed: 04.10.2025) ↑
44. United Arab Emirates. Federal Decree-Law No. 45 of 2021 on the Protection of Personal Data. — Abu Dhabi, 2021. Available at: https://www.dataprotection.gov.ae (accessed: 04.10.2025) ↑
45. United Arab Emirates. Federal Decree-Law No. 46 of 2021 on the Establishment of Digital Government. — Abu Dhabi, 2021. Available at: https://laws.uaecabinet.ae (accessed: 04.10.2025). ↑
46. Republic of India. The Digital Personal Data Protection Act, 2023 (No. 22 of 2023). — New Delhi: MeitY, 2023. Available at: https://www.meity.gov.in/data-protection-framework (accessed: 04.10.2025). ↑
47. Republic of India. Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 (as amended 2023).— New Delhi: MeitY, 2023.
    Available at: https://www.meity.gov.in/writereaddata/files/Intermediary_Guidelines_and_Digital_Media_Ethics_Code_Rules_2021.pdf (accessed: 04.10.2025). ↑
48. NITI Aayog & MeitY. National Strategy for Artificial Intelligence #AIforAll; Responsible AI for Social Empowerment 2024 (IndiaAI Mission Documents). — New Delhi, 2024–2025. Available at: https://indiaai.gov.in (accessed: 04.10.2025).
    ↑
49. Anders G. The Obsolescence of Man. Vol. 1. Munich: C.H. Beck, 1956. — p. 22. (Russian translation: Obsolescence of a person. Vol. 1. Moscow: RIPOL, 2002. — P. 41–42.) ↑
50. Lem S. Summa Technologiae. Warsaw: Wydawnictwo Literackie, 1964. — p. 43. (Russian translation: Summa tekhnologii. Moscow: Mir, 1968. — P. 52.) ↑
51. Kant I. Groundwork of the Metaphysics of Morals. Cambridge: Cambridge University Press, 1998. — p. 436 (Akademie Ed.). (Russian translation: Fundamentals of the Metaphysics of Morals // Soch. in 6 vols. Vol. 4. Moscow: Nauka, 1965. — P. 249.) ↑
52. Loza de Siles E. Slavery.AI // Washington and Lee Journal of Civil Rights & Social Justice. 2024. Vol. 30, iss. 2. P.214 URL: https://scholarlycommons.law.wlu.edu/crsj/vol30/iss2/4/ (accessed: 03.10.2025). ↑
53. Loza de Siles E. Slave.io // SSRN Preprint. 2021, P.7. Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4056443 (accessed: 03.10.2025). ↑