Artificial Intelligence in Art: An Amoral Subject of Law

What Artificial Intelligence Can Pose this Type of Problem

In the first place, it is necessary to differentiate that, within the vast world of AI, only a certain specific type is the one that can give rise to these problems, since the rest do not even enter into such a question nor do they escape from being another mere object of Law and of social reality. It is necessary to differentiate that there are, broadly speaking, types of weak AI and strong types. Regarding these qualities, Grandhi collects them in a differentiation of different levels in two blocks. First, understand that there are the weak-type and strong-type AIs on the one hand, and then the symbolic and non-symbolic learning ones [8].

Within the first block, in the case of weak-type AIs, their main characteristic is to perform specific tasks, not to be reactive, which means that it does not act autonomously, it is not flexible, and therefore it depends on the human programming through imitation, since it does not have any capacity for reasoning or learning. In the second case, the strong type, it does have the ability to carry out tasks that are usually covered by the human intellect. This allows flexibility of action for problem solving, where here it does participate proactively, with qualities of self-programming through techniques such as feedback through complex neural networks.

Then, that at that programming level there are supervised and unsupervised algorithms; In addition, the processes can be of a symbolic nature, where there are disciplines from machine learning, deep learning, neural networks, self-learning. Within those of a symbolic and non-symbolic type, those of the first group focus above all on the fields of mathematical logic, which in turn is inspired by systems of rules, as well as representations of knowledge. On the other hand, unlike this type, non-symbolic ones focus on systems more similar to simulations or emulations of human behavior, such as the process in which the nervous system works, such as the neural network system, or self- learning. that occurs in living organisms. Between these two groups, it is the second that embodies the most “ambitious” branch in that it seeks to simulate or resemble, at least, the different systems and processes that operate in the brain. In a few words, the AI that can provoke these dilemmas is only the strong type, unsupervised [9], of a non-symbolic nature and, furthermore, with a greater peculiarity, the resolution process between the input and the output of information where the evolution of the IA escapes the control and supervision of any third party [6], such as the programmer, that is, the concept of “black box”: In essence, this means that algorithms cannot provide a detailed explanation of how they arrive at a given result. That is, it cannot be established how the AI system evaluates and weighs the data and information it processes. That is why we speak of “black boxes”. A computational tool in which one understands the input data and the results, but does not understand the underlying procedure, is called a black box system. Here the code is inscrutable, because the program “evolves” and human beings cannot understand the process that the programming followed to achieve a given solution.3 The main characteristic lies in the absence of the human factor in the processes carried out by the AI, which implies that the possible definition of “autonomy” that arises is in a negative conception of the term: we understand that an AI can be autonomous or independent as long as no human can intervene or supervise the process developed by the AI, unlike the autonomy of the individual, which is from a positive conception, where its autonomy emanates from itself. Another characteristic is that all the processes described consist of the emulation and simulation of the human process, that is, AI can only cover the technical field even in the volitional factor of the human being, it can only copy, simulate, emulate the decision-making capacity, the corpus mechanicum of the human being, however, the animus vivendi behind the technical factor and which supposes the origin of the indissoluble incentive of the person escapes AI since the intellectual translation of the supra -technical qualities of making a decision or not , to create or not a work are beyond the scope of all programming. That it can emulate the human being does not mean that it is human.

From Object to Subject: An Axiological Alchemy of Legal Personality

María José Santos González, coordinator of the Legal Department of the National Institute of Cybersecurity of Spain, commented in the study “Legal regulation of robotics and artificial intelligence: challenges for the future”, these types of questions about studies by both the Gartner consultancy where, In theory, in 2020 we would talk more with robots or AIs than with people, like our own partners, for example. Also, it warned about the concern of several

3  Corvalán, J. G., 2017. La primera inteligencia artificial predictiva al servicio de la Justicia: Prometea. LA LEY, p. 3.

psychologists that this would mean blurring the sometimes apparently thin line between reality and fiction, the natural versus the artificial. However, can the human mind be equated with the artificial one? Are we not obsessed with the corpus mechanicum instead of the corpus mysticum hidden behind the facade of technique? As Santos González asks, “can they commit crimes? Should a robot be judged for its actions the same as a person and at the same level for being artificially intelligent, even though it lacks real emotions and feelings? How can we control the robot? [10].

Or, on the contrary, are we not granting with these issues relations in the sphere of ethics, morality, virtue, which have been and are enforceable and adjudicable to the subjects of Law due to the intrinsically subjective nature of such matter? A chair cannot be required to behave well, however, a natural person can and, on the other hand, the requirement to a legal person lies in the indirect requirement to the natural persons that make up the company (even if the participation is in turn divided into other companies, since at the end of the chain there will be at least one person in charge behind it and who will take over the deliberative processes). When good behavior is required of a company, it is really required of the people who make it up. The telos of the ethical and moral demands concern the behavior of the subjects, not of the objects, since in this case they are the people responsible for the thing. Otherwise, we would be faced with a requirement of ethics of movable property, as if a movable property could comply with a code of conduct and exemplify moral behavior by itself.

Therefore, if we do not require a chair to behave well; if one could not present or be proposed for elections and hold the right to active suffrage or be in charge of employees or, in short, those related to fundamental rights, why would an AI or an intelligent robot be able to do it or, above all, why the effort that it can or should? Can an AI or intelligent robot really be considered the author of an artistic, scientific or literary work (art. 2 LPI and 10 LPI) and that it enjoys a “digital originality” [11], and a “computational creativity” understood? According to López de Mántaras as “the study of software development that presents a behavior that would be considered creative in human beings. This creative software can be used for anonymous tasks such as inventing mathematical theories, writing poetry, painting pictures, and composing music” [12]? And an “algorithmic authorship” [13]? Or, even, if it must have a legal personality, be an “electronic person” which, in the words of Valente: Electronic personality would mean considering robots as a legal person who has certain rights and obligations of a merely instrumental nature for a specific economic interest of a human being [14].4

4  Valente, L. A., 2019. La persona electrónica. Anales De La Facultad De

Is the AI going to hold rights to privacy, honor and its own image? Will it enjoy a digital dignity in such a hypothesis? All these scenarios when, in the best of cases or the most revolutionary, an amoral subject could be achieved, at best, and in the most plausible event, an intermediate position between the object and the subject could be considered (pp. 181-182) [1], that is, that something can escape from the mere consideration of a thing or piece of furniture due to the fact that, unlike the so-called chair, an AI has the functionality to act in fields that require a deliberative and decision-making process, the volitional function of making decisions (probabilistic in the background). However, that an AI can perform techniques in disciplines of the human person does not imply that it is a person (and much less human), but not because the claim of such consideration emanates from the AI, but because it can have a practical utility to solve certain highly complex situations. That an AI can produce a painting, a song, a poem does not imply that it has the animus vivendi, the intrinsic creative incentive of the human being, which is not programmable, since it cannot be “the author of something that only a physical person can do “. or natural” [15]. On a technical level it can be an acceptable work, even magnificent, on a technical level, but why would it not be protected that an AI paints a yellow dot on a canvas (a reductionist example of an abstract painting) and if the human being does it, yes?

The conception of originality, creativity, authorship or projected person is legitimized from a negative definition, since it is the non-human intervention that leads us to be able to state that it is the AI that intervenes in an “independent” way, but any degree is enough. of human intervention to then lack such a foundation, unlike a human being, whose legitimacy is affirmative, part of the subject itself and even with any type or degree of intervention by a third party, it does not lose an iota of its ownership or originality , a requirement of originality which, in the words of Bercovitz Rodríguez-Cano, “is the essential requirement for a creation to be considered a work” [16].

On the subject understood as the one who can claim it, according to the premises of the North American WN Hohfeld [17]5, Sauca highlights in the modalities of subjective right (what refers to claim, freedom, power and immunity, categories of WN Hohfeld) in terms of power, which “is in the legal situation of power one who has a type of subjective right such that he can modify his legal relationships against another” [18]. And in such a matter, an AI or an intelligent robot are not capable of modifying legal relations with Ciencias Juridicas Y Sociales De La Universidad Nacional De La Plat, Issue 49(01), p. 13.

5 Véase Hohfeld, W. N., 1913. Conceptos jurídicos fundamentales. México: Fontamara, pp. 45-87.

other third parties, they can interact, just like a sentence, an administrative act, a contract interacts and for that reason we are not in a hurry to give it a legal personality so that they are subjects of Law, unless you want to go beyond fiction and conceive of a contract such as a supermarket purchase as a proper subject of Law.

In many cases, going beyond the field of reality through fiction through a theoretical reality, a hypothetical reality, within the Philosophy of Law, and of Law in general, the fictio legis that inspires every norm and that is the backdrop enters the apparent neutrality or impartiality of the norm and the axiological factor that is hidden behind it. Can an AI or a robot claim rights? Could I become a subject of Law through the claim or action of claiming them? Are we not, in truth, the ones who project such aspirations? Doesn’t the problem lie in the fact that within the personality that they can sustain, it is nothing more than another regulatory artifice, a legal invention that is extrapolated from that formal reality to social reality?

The most imminent example is that of legal persons. A company, no matter how much effort and legal and philosophical engineering it projects, does not escape being a ruse as a vehicular instrument that serves the human being in their contractual relationships. A company exists in its bylaws, in the corporate members that make it up, however, no one can have a coffee with a load of statutory paperwork, nor can one with the abstract conceptualization of the company: you need to be a person, made of flesh and bone, with whom you agree such a summons, with whom you order such a decaffeinated snack or not, with hot or cold milk. Similarly, with AI, which is and will not cease to be a piece of furniture, no matter how many logical pirouettes it may take, where the demands of ethics, principles, values, behavior or, on the contrary, the projection of a piece of furniture as subject capable of claiming rights are not feasible, just as no “company” as an abstract entity presented itself to claim any right, but the (human) person who represented it did.

So, regarding the electronic person (a proposal that was formulated by the European Parliament and which was not without controversy), it would be to create one more modality than that of a natural person and a legal person. On the other hand, the main problem with these proposals is that, although behind the legal entity there will be natural persons who provide the moral subject of the company or company, on the other hand, in AI it is proposed without a human subject behind, because then it would not be this debate in question (that individual would be responsible and subject to Law, as has been done to date): from the proposals to provide legal personality to AIs and intelligent robots, it is proposed with both the economic and morality of the subject of Law to an entity that, in the best of cases, does not manage to become more than an amoral entity, moreover, it is just this intrinsic amorality to AI when applied in ethical fields where the human being who deliberated problems have occurred , since the human being is not a computer nor an algorithm is a person [19]. An entity that does not understand what is good or evil (within a consensual social morality and in accordance with Law) is put to decide as if it were in equal conditions to do so as a human being.

Unlike the proposals of an electronic personality, it would be better to understand an intermediate category between object and subject, where it is accepted that it is true that an AI acts in fields that could previously be exclusive to the human being (a chair did not have the deliberative function to determine if the economic benefit is accepted or not to a citizen and an IA yes), it does not cease in essence to be a movable asset capable within the corpus mechanicum, that is, within the technical field of those deliberative-resolution processes where it participates, interacts in a way somewhat comparable to how a person would and, therefore, transcends to a small degree the situation of the thing or as an object of Law, but that being exempt from all corpus mysticum , from all incentive will (other than to the initiative, which is programmable) and of all anymus vivendi , it does not reach being a full subject of Law, since it will not escape being an amoral subject. And, therefore, what it needs is not to frame it in our same parameters, but a sui generis right 6, if anything [20].

So, this intermediate classification understands that, in the most advanced of cases, what could be called a cyber- humanoid personality would be between a supra-object and a quasi-subject , that is, a subject limited only in the eyes of the Law (not in the of social reality) to be an economic subject with which to respond civilly for the damages caused and, therefore, to be able to respond to such demands, the ability to safeguard a monetary heritage that thus satisfies the claims of the third party, not like the moral subject that is a natural person or a legal person indirectly by the individuals that compose it, why? Well, due to the human absence in the deliberative processes in which the AI is involved. Just as it would not be fair due to the distant or almost non- existent causal link between that process unknowable to the programmer that is the “black box” making him responsible for the decisions that an AI has made, for example, with racist biases or other discriminatory nature. Programmer who can respect all the civil rights of every individual, even being the greatest defender or even having such identity characteristics and who neither programmed nor could prevent nor knows the slightest in making such decisions (this extends to the

6 Véase Ríos Ruiz, W. R., 2001. Los sistemas de inteligencia artificial y la propiedad intelectual de las obras creadas, producidas o generadas mediante ordenador. La Propiedad inmaterial, 13, p. 11.

user of the AI), it would be unfair to assign responsibility for such acts. On the contrary, it is equally unfair to assume that the effects of an action on a third party are exempt from their responsibility.

With this, given the absence of the human factor attributable to the harmfulness of the act and given the prudence of entrenching in a vacatio legis , the scenario of that intermediate category can be conceived from a legal projection, however, this should not be extrapolated situation from an anthropological perspective to the social reality, since it would be to confuse the principle of ubi lex non distinguit, nec nos distinguit due to the human factor, that is, if for a fact we cannot distinguish who causes it, to reason that the possible causes are analogous or similar. And this is what happens, for example, with creation and authorship. In other words, although in the legal field, due to the very artifice that defines it, an intermediate personality such as the one proposed would be feasible as a vehicular instrument, it does not escape being a legal trick, without forgetting that it would be for the exceptional situations of the most advanced.

An Author Object?: Is an AI the Author of a Work by the Fact of Being Able to Originate It?

The proposal to consider an AI as an author implies accepting a type of process where an object can “evolve” into a subject, since the fact that no matter how advanced the AI is or becomes, its source code, should not be neglected. the algorithms that integrate it, are protected as an object through Intellectual Property as a work, since its nature is analogous to that of any other variety of protectable creation. It is, in essence, an object, because that is how it is created, unlike a natural person, who is born as a subject and his personality is recognized from birth (including the protection of his rights and interests before in the figure of the unborn child ) and to the legal person that is born at the moment in which it is formalized in accordance with the Law and its requirements: both cases, from their first moment, are born as subjects of Law and there is no way to ensure or violate such condition, instead , the AI is born and is conferred from its origin as an object and, even when it “creates” a work, no matter how independent and particular that fact may be, it does so from its legal status as an object, a work, a movable asset.

Then, through an inductive reading, considering that an AI from meritorious criteria for the result, as a kind of iuris tantum, author or with copyright, supposes a normative ubiquity of the same legal right. The AI is registered and protected as an object, but its results or works are protected with the AI as the subject. How would Law and Philosophy deal with an entity affected by this duplicity? Like Shrödinger’s cat, the AI would thus be being an object and a subject at the same time on two different legal levels.

This means stating that an AI can hold rights that are not only subjective but fundamental, since in truth the freedom of creation is a subgenre of freedom of expression. Among the most essential requirements to be an author or inventor (depending on whether, in the first case, we are in Intellectual Property or, in the second, in Industrial Property, although since the WIPO and most Anglo-Saxon regulations, Property law combines the two disciplines), highlights from a subjective point of view the factor of human ingenuity, human inventiveness and human personality that is reflected on the work or creation and, from an objective scope, the need to enjoy legal personality in order to hold or direct ownership of the rights, that is, the legal author or, in any case, being the material author.

Moreover, from this objective distinction inspired by Locke on his conception of property and the theory of fruits, the legal artifice of being a legal person is the owner of rights as employer and the material author is the employee who creates the property. work (it is the figure of Work Made For Hire). Or, from the second utilitarian view of Locke’s approaches in this field on the principle of efficiency and “best use”, on which it is argued that recognizing an AI or an intelligent robot of subjective and moral rights such as those of authorship is efficient and practical, not only for Law, but for social reality, it implies deconceptualizing or “vulgarizing” the qualities of the concepts of originality, creativity and authorship of its most vital elements, since they are intrinsic characteristics of the corpus mysticum of what is protected and what the work represents, which is the personality of the author, the imprint, the human ingenuity that is a different element from the technical corpus mechanicum of the work. Previously, the example of abstract art was mentioned, which focuses on that subjective quality, personality, and if faced with an abstract painting that would be protectable and its protection would be understood, this would be unthinkable if that same painting were executed by an AI, because the reason for reflecting on the work done by AI is, in truth, due to the meritorious factor that the resulting work at a technical level is similar to the human one, that is, that it is considered that the result can be “original ” because it emulates at a technical and meritorious level a result that could have been generated by a person. For this reason, it is alarming when it is found that in the Chinese judgment of the Court of Shenzen, of the province of Guandong, a work produced by an AI is protected with the same copyright that a human author holds, by finding a logical, coherent structure and a some degree of originality [21].

What originality are you referring to? Well, it is not questioned that a work must be produced ex nihilo, nor is it alluded to that the creation must be ex novo, since any author or inventor is nourished by the entire cultural and intellectual heritage to form his own creative personality [22] and thus create with the unavoidable contribution of his personality and human ingenuity and through technique. However, of the two essential elements of the author, finding that a work produced by AI can meet one of them (which at a technical level is somewhat similar), one should not rush through a retrospective reading to conclude that it meets the sum of all the requirements, since the anymus, the creative incentive does not exist in AI, it is in any case the mere essay- probabilistic mirage of an algorithmic combat within its input and output process.

With this, it is possible in collation to the previous section, applying the instrumental creation of the intermediate personality, to understand that in the best of cases we are facing a quasi-authorship as limited or secondary authorship where it can be accepted that the corpus mechanicum as one of it fulfills the requirements for authorship, in the sense that the result, at a technical level, is protectable by the legal interest that concerns it, although it lacks any anymus since, being in the best and most fictitious of cases, an amoral subject, neither can nor does it have a corpus mysticum or a moral factor or a personality that permeates the work. Therefore, being at most in that quasi-authorship that the material and technical author of the work is proposed, he is not a moral author nor can he be of the result. Just as little spirit can be expected to be born from the four legs of a chair, so little creative spirit can be expected to be born from the frantic shuffling of algorithms and source codes, ones and zeros.

In addition, because among the plausible philosophical- legal approaches for the options of being an AI full subject, full author, it is based on a dehumanized utilitarianism since it is empowered in an efficiency that contradicts itself, because if the “individual ” to which you are recognizing some subjective rights is an entity incapable of claiming such rights or even assessing an interest and incentive in its claim since it is an amoral subject, it is, in truth, more inefficient to provide it with that capacity, where it should be well Bearing in mind that, by doing so, such right becomes a ius prohibendi against third parties, that entity that is an amoral subject holding moral rights, is in itself totally incapable of claiming them, therefore, if there is the possibility of separating and to break off that capacity so that either by legal representation, or by how the figure of the third party is adopted (although an “algorithmic” guardianship or conservatorship falls on the same matter, what is the m of an AI?), this possibility vanishes, distorted in a hypothetical subject that holds a right that in ius prohibendi is the only one qualified to exercise it, it turns out that he cannot, compared to a third party who could. In

summary, as Saiz García rightly says, no matter how much effort is made, “the machines do not have, unlike human beings (at least still and almost all), conscience or affection that could be influenced by the exclusive protection of your effort [23].” With which, the moral protections by which this type of rights emerge and that are sustained in that exclusive sense of the effort of our conscience and sensitivity would lack any axiological sense with the partura to entities that only comply, in the best of cases , the technical requirement, the corpus mechanicu , excluding not only an essential element, but perhaps the most fundamental of both, the anymus of the subject, the corpus mysticum of the work, one of the differentiating and indissoluble characteristics of the human being, of the full subject of Law.

Conclusions: Object and Subject? The Box Uncovered

In conclusion, according to everything exposed in this work, it can be concluded in the first place that no matter how extensive the rain of anecdotal and contradictory events that take place in reality, the Philosophy of Law should not be clouded by the fires of artifice and confuse them with the protean fire by which it emanates. From the nature of the very fact that an AI supposes, which is that of an object as the work of an individual, the programmer, who will protect such work in accordance, this yes, to his copyright, devirtualize by the meaning of the matter to “ promoting it” to the category of subject of Law at a level, moreover, that would be comparable to that of the human being, either as a natural person, or as a member of the legal entity, means confusing or reducing the complexity and terminological diversity of too many concepts , which are not simple in themselves. However, although it is true that the AI has the ability to take part in deliberative processes that normally depended on the human factor, this does not mean that they turn it into a human, but rather that one of the properties elaborated in that object is that of being able to carry perform such functions.

Otherwise, to confuse the telos of an object in this way is to conceive that a machine capable of performing surgical operations must therefore be considered a surgeon; that an AI system for predictive data within the public administration must be understood as a public official; that, in our case, an algorithm, AI or intelligent robot, due to its ability to produce a work, must hold a copyright, therefore, when it is also already protected as the work of the programmer who created it. In this line, without the least cases and very restricted to a specific type of AI that may be capable of developing such types of faculties, however, even in the best of cases, it has been possible to clarify that not in the most advanced and hypothetical does not transcend beyond an amoral subject in the best of results, since assigning him an ethical or moral aptitude is accepting that it can give an ethics of movable property, that movable property can comply with an ethical attitude or behavior or , even more, that an object has its own morality, something impossible beyond the field of mythology and literature.

Thus, as in Shrödinger’s theoretical experiment, the cat is inside the box with a bowl of poison, and until the lid is uncovered, the cat is alive and dead at the same time. If, in this case, the “black box” of the algorithmic process is uncovered and discover if it is an object and a subject, the reality of its nature (and not only legal) emerges: it is an object. Not an object-author. Not an object-subject. An object-object that is regulated and protected as a work in accordance with copyright, a right held, precisely, by its programmer, who is indeed a subject [24].