legal personhood of artificial intelligence

July 12th, 2023

introduction: journey of ‘the person’

artificial intelligence, which we can describe as the trendiest topic of recent times, interacts with the law through many different channels. as a result of such interaction, various questions arise about where artificial intelligence assets can be positioned within the framework of the law. the question that i think is one of them is about whether artificial intelligence will have a legal personhood. to find the answer to this question, first, it is necessary to examine the existing personhood state and then to consider the situation regarding artificial intelligence and the personhoods of non-human beings. it is hoped that ideas about the legal personhood of artificial intelligence will be formed thanks to this situation pattern.

first, since it is necessary to explain the existing situation, it would be a correct choice to talk about the various theories based on the legal personhood of the human being.

here we can talk about a journey that goes towards modern times. when we look at the beginning of this journey, we come across the name boethius. boethius gives us a definition in his contra eutychen et nestorium: “a person is an individual substance of rational nature”. looking at the definition here, it is seen that there is a sign of individuality and besides this, it is emphasized that belonging to a rational nature. such a general definition would certainly be open to interpretation, but it does not seem possible for a substance that does not exist in nature or that does not belong to nature to be a person. this approach commands us to include people and other natural beings in the person cluster, and this approach is also known as “the classical conception of the person”.[1]

when we take another step in our journey, this time a new name emerges: john locke. locke narrows the framework in the light of the fact that a person is a being who can think and has intelligence, with his thoughts in his essays concerning human understanding. the act of thinking is not seen as a momentary act, it draws attention to the fact that the beings who can reflect themselves in different places and times can be called a person. contrary to boethius, a distinction emerges not based on whether it exists in nature or not, but also on the abilities it possesses after it exists, and at the core of this distinction is the ability to think. for this reason, this approach is known as “the psychological conception of the person” and drags us to another stage in our journey.[2]

in our new phase, this time we come across the name immanuel kant, and he defines his metaphysics of morals: “a person is a subject whose actions may be imputed to him”. here, the adventure of identification goes one step further and beyond the ability to think, there is a transition to the level of whether the being takes responsibility for the actions it has done. if the entity can act by thinking about the before and especially the aftermath of any action it has created with a sense of responsibility, it means that it can be referred to as a person at that point. this approach is also known as “the ethical conception of the person” and as the name suggests, it has a moral approach. the awareness of responsibility and the actions of the beings accordingly create the need for progress in the light of some abstract principles and these principles come to the fore with their moral dimension.[3]

to summarize before moving on to modern theories, for an entity to be accepted as a person, it must first be a substance belonging to nature, then be a thinking entity, and finally be aware of its actions. as a result of this complementary approach, when the beings on earth are looked at, the human being draws attention as the only being that can meet the mentioned conditions -at least as we know. in other words, after the above-mentioned approaches to personhood, it is concluded that only people can be referred to as a person.

as we continue our journey in the historical process, we come across more critical and modern approaches and new names that can be associated with these approaches. first, if we talk about the “descriptive theory of the person”, peter singer argues that a person should have some skills. these abilities can be listed as feeling and understanding, self-awareness, autonomy, imagining oneself in the future, but they are not limited to those listed. however, when interpreted from a critical point of view, it is seen that not every human being, who we previously considered as a sufficient entity to be a person, can have the abilities. for example, the dilemma occurs that a human being in a coma cannot have these abilities, but on the contrary, some animals cannot have all these abilities. in such a case, if personhood is attributed to the human being in a coma, the same approach should be applied to the animal in question.[4]

however, hugo engelhardt clarifies this situation with the following lines: “not all people are equal. […] not all people are persons. not all people are conscious, understanding, and able to praise or criticize something. a fetus, a newborn, the mentally handicapped, those in a deep coma – are examples of people who are nonpersons.” as engelhardt also stated, it would not be correct to refer to every human being direct as a person. at such a point, it is necessary to turn to singer and listen to his criticism of the distinction made. singer states that the distinction between human and non-human in ethics and law should be abandoned because all people are not equal and not all people are persons, instead a distinction should be made as person and non-person. from a different point of view, if a distinction such as a person and w non-person is accepted, a possibility arises that they cannot be filled only with people, and the thoughts that the person is seen as equivalent to the human being are kind of deconstructed.[5]

looking at another approach, namely “the axiological conception of the person”, it is seen that the focus is on the ethical value of the person. ethical value and ontological priority are considered sufficient to define the personhood of the entity in question, and contrary to the claims of singer and engelhardt, the idea that every human is not equal is not considered, and any mental disorder or developmental stage of the body is not considered important. just being a human is seen as enough to acquire the personhood and supports the idea that human being stands in a superior place among all beings in the world order. in a much simpler expression, a sanctified entity is mentioned, and the understanding is conveyed that only this being will have personhood.[6]

personal subjects of law

along with having personhood, it is necessary to be a subject of law at the same time to reach a competitive position within the law. it is also clear that being human alone is not enough to be the subject of law.

when the relevant civil law provisions are examined in general, it is seen that each person has the legal capacity with his/her birth. of course, this approach may contain additional conditions and/or different perspectives depending on the country differences, but when generalization is made, the intersection set is obtained in this way. however, there are still some question marks about the fact that every human with legal capacity is a subject of law. as mentioned before, not all people are equal and the differences between them also affect the law. to remind again, according to the descriptive theory, a person should have abilities such as feeling, understanding, and self-awareness. if there is a problem with these abilities, we can still talk about a person in the legal sense, but we cannot talk so confidently about being a subject of law.[7]

according to the provisions of the civil law, we determine the persons who are the subjects of the law according to some qualifications in addition to the general abilities in question. for example, conditions such as being over a certain age and being in good mental health reveal the differences between the person and the subject of law at this point.

at this point, the confusion, and the feelings about how problematic the concept of person is revealed. the content of the concept of person, which has been tried to be explained from the very beginning of this study, the limitations are drawn to this concept, and how the concept in question will be evaluated in the law is as follows: while the person is a human being who is natural and has some abilities, the subject of the law is a group of people who can use and fully understands general abilities.

contrary to the general definition made from the very beginning to the point we came to, it is seen how a small group reduction is. from this point of view, especially from the point of view of the law, it is deduced that people adopt an approach that considers themselves superior to other beings. in connection with this, legal systems take their place in the order as systems managed by people for the sake of people’s interests.

in a system that works for the interests of people, various subjects other than people can take place, and these subjects are referred to as human communities. subjects such as companies, municipalities, associations, and foundations, which are created by people and accepted as subjects within the system, and most importantly, not belonging to nature, have been created over time. although it is known that these subjects are created by people coming together, it should not be forgotten that they are artificial, and it should be noted that this development opens space for us to be artifact persons. there is a sad but true thing that people’s interests are at the forefront for this artificial personhood to come into existence.

finally, it should be stated that the legal system is a structure in which things exist as well as subjects. here, personhood is defined as the capacity to have rights and responsibilities, and those other than the subjects who have such a capacity are described as the object of these rights and responsibilities. this duality, which is expressed as persons and things, is connected on the one hand, but at the same time intertwined with each other. currently, when it comes to rights and responsibilities, we come across people and human communities, but it is known that not every person is included in this field. other representatives, that is, people, are assigned to save the rights and responsibilities of certain people (for example, minors, mentally ill people, etc.). from another point of view, when we look at the ongoing discussions about animals, these beings, which are accepted as objects of rights and responsibilities in some legal systems, are the subjects of law in some countries, and this change is progressing day by day. in such a case, non-human legal subjects come to mind.[8]

non-personal subjects of law

since roman law, a relationship has been established between being the subject of law and being human. this relationship is not bidirectional, that is, while every human being is accepted as a subject of law, not every subject of law is human. the difference between being the subject of law or being a thing appears at the point of having rights and freedoms, and this distinction is also used in the definition of the concepts. accordingly, a person is defined as someone who can save all kinds of rights and freedoms. on the contrary, it is possible to define non-personal law subjects who fall into the category of things as those who can exercise only a right, such as legal recognition, not all kinds of things.[9]

the non-personal subjects’ approach can only be applied to emotional animals. even though it is still controversial, animals that we accept to feel, and think are both involved in people’s lives in a way and can also enter legal systems with the help of this. however, this approach does not apply to non-emotional living organisms, plants, and countless living beings apart from emotional animals.[10]

from another point of view, the non-personal legal subject approach is not considered sufficient in terms of its application to artificial intelligence assets. the reason why it is not considered sufficient here is that artificial intelligent beings cannot have consciousness and emotions now and soon. however, it should not be forgotten that not all artificial intelligence assets are at the same level as each other, and some of them have very complex structures. at such a point, generalizing all artificial intelligence assets into the machine category is also a problematic approach. it is known that especially autonomous artificial intelligence assets take themselves to a very different level by getting rid of the data given by their creator, and such assets should be positioned in the legal system, at least in terms of their responsibilities.[11]

the anthropocentric attitude accepted today is being questioned from the ethical framework. as mentioned above, it is obvious that a post-humanistic perspective, in which non-human living organisms and artificial intelligence entities are included, is now inevitable. within the post-humanistic perspective, it is stated that not only people, but also all other beings should be ethical subjects, that is, they cease to be objects of people, but the set of entities that can be considered as ethical subjects differ according to different approaches. at this point, two basic approaches can be mentioned:[12]

first, according to the environmentalist ethics approach – also called green ethics – different elements in nature are included in the ethical cluster. this approach also has two subsets within itself. according to the light green ethical approach, it is stated that emotional beings, that is, beings who can feel pain and who can suffer, should be accepted as ethical subjects and therefore legal subjects. according to a presupposition of this approach, non-living beings cannot feel pain anyway, and therefore only a part of animals is considered in this approach. according to the dark green ethical approach, which is another, no limitation is foreseen for animals that suffer only, but all beings in nature are accepted as ethical and legal subjects. in addition to animals in the light green ethical approach, others in the animal kingdom, forests, rivers, living organisms are included in this approach.[13]

secondly, according to the artificial intelligence ethics approach – which is also called silicon ethics – all kinds of entities that emerge because of different scientific developments are included in the ethical cluster. the ideas of those who advocate this approach are based on three generally accepted statements. according to this, (i) brain processes are informational processes, (ii) brain processes are epistemically accessible, (iii) every cognizable informational process may be emulated/reconstructed in a computer. with the acceptance of this statement, it is scrutinized that the ethical status is related to the characteristics of the informational process and that the material substrate does not have a say in determining the status in question. it is stated that all kinds of informational processes are qualified enough to reveal absolute beings to be accepted as ethical subjects, and this idea is defined as info-centrism in some academic works.[14]

however, when we look at today’s ethical approaches, it is seen that is still anthropocentric, and on the other hand, various solutions are tried to be found according to this feature. in this approach, artificial intelligence entities are accepted as electronic slaves, and artificial entities created by humans are treated as something rather than a subject. this attitude does not present a status very different from the status of slaves that we saw in the roman period. the concept of peculium, which was revealed by romans, finds its counterpart in the current anthropocentric approach.

it is possible to define peculium as follows: “the property held by a person (as a wife, child, slave) under the potestas, manus or mancipium of another as his own private property either by the permission of the paterfamilias or master or by the rules of law but becoming with certain exceptions the property of the paterfamilias or master at his pleasure.” similarly, a structure in which limited responsibilities of artificial intelligence assets are accepted is considered. contrary to a model where all responsibility is left to the creator, it is thought that the creator is not fully responsible for the actions of artificial intelligence entities, but in a limited way. because, especially autonomous artificial intelligence, as mentioned before, can make a different creation by using the data they receive from their creator and can reach places beyond what the creator demands. in such a case, holding the creative legal subject fully responsible would be unfair for that person at this point. as such, the currently accepted anthropocentric approach is stripped away from the conditioning of subjectivity such as the ability to suffer or the capacity of consciousness.[15]

however, we know that it is possible to come up with negative situations as well as positive ones. although these negativities may be resolved in time, they currently leave some question marks in mind. although we will talk about a few sample questions and problems here, the number of them can of course be increased.

the first of these question marks are about the status of autonomous weapons during the war. although some of the currently used unmanned vehicles are managed in the form of remote control, some of them can make autonomous decisions and act. during the war, concerns are also emerging based on whether the autonomous vehicles in question can discriminate, and the situation of civilians in the warring country, where there is a lot of effort for protection under humanitarian law, is going to a more dangerous place. as a result of the decisions made by autonomous artificial intelligence entities, people’s lives are at stake, and leaving such an important situation to their decision becomes controversial. it is also unclear who bears the responsibility for a possible wrongful act destined for the lives of civilians. since autonomous artificial intelligence entities act independently of the will of both the creator and the person responsible at that moment, who and what will be sanctioned because of the action taken becomes an inextricable situation.[16]

secondly, even though people’s lives are not on their end, software used in commercial areas also has the potential to violate people’s rights and freedoms. it is not acceptable in an anthropocentric approach for the software in question to make decisions that cannot be predicted by people with self-learning complex algorithms. in addition, the software in question acts in different ways to reach the target given to them and strives to reach the target most shortly by improving their algorithms. the actions of this software on the way to the target may also not be predictable or not appropriate both legally and ethically. although there is no question that it violated any rights when it reached the goal, what it did while moving on that path will be open to discussion.[17]

the subjectivity scenario can be used to solve the problems that may arise with the use of artificial intelligence in both the war situation mentioned above and the situation in commercial transactions. for this scenario, first, it should be accepted that artificial intelligence assets are not only harmful but also beneficial for humans. from reducing the casualties during the war by reducing the number of people fighting, to ending the war in the short term; different benefits can be counted from accelerating the mobility in commercial transactions to obtaining economic vitality in the world order. however, it emerges because of the unpredictable actions of artificial intelligence, which is seen as harm here, and even if such an action does not occur, the possibility inevitably makes one think. for this reason, in the subjectivity scenario, it is recommended to organize artificial electronic personhood, like companies, and to transfer the actions that are not within the responsibilities of people related to artificial intelligence, such as the creator or user, to this regulated personhood. of course, to organize such electronic personhood, some conditions will be put forward and artificial intelligence entities with certain criteria will be recognized as persons. in this case, some of the artificial intelligence assets will be accepted as the subject of law, and the rest will continue to be an object.

conclusion: ai in the future

when we conclude point, it is possible to see that some experiments have been made on artificial intelligence entities that can be the subject of law. the european parliament states in its “resolution on civil law rules on robotics” in 2017: “creating a specific legal status for robots in the long run, so that at least the most sophisticated autonomous robots could be established as having the status of electronic persons responsible for making good any damage they may cause, and possibly applying electronic personality to cases where robots make autonomous decisions or otherwise interact with third parties independently”.

although this decision of the european parliament causes controversy, it must be admitted that; all ai assets and machines are products. although autonomy makes it independent from human control, it is not possible to talk about any responsibility gap. in any case, the responsibility of the creator and/or user of the artificial intelligence can be compensated and/or sanctioned if necessary. however, these people need to be considered. when such an attitude is taken, the impression is created that it is aimed to punish people for the existence based on artificial intelligence that they have created. however, on the other hand, considering that the level of legal personhood itself is artificial, it should not be forgotten that it was created for purposes such as ensuring coordination between the parties, sharing responsibility, and differentiating taxation. accordingly, attributing personhood to artificial intelligence assets in a basic sense will fulfill such purposes and will not prevent decision-making according to concrete events.

at this point, it is appropriate to conclude the study by mentioning a precedent. limited liability companies (llcs) can be taken as an example, as suggested by american academics regarding the electronic personhoods of artificial intelligence entities. we can explain what these companies are as follows: “an llc provides its owner or owners with limited liability. this means that means you—the llc owner—are generally not personally liable for any debts incurred by your llc business or most business-related lawsuits. because you are not personally liable, creditors or people who file lawsuits against your llc can not collect against your personal assets like your personal bank accounts, personal car, or home. llcs ordinarily provide their owners with pass-through taxation. the profits (or losses) the business incurs pass through the business to the owner’s personal tax return. such profits are taxed at the owner’s personal tax rates.”[18]

with the hope of finding solutions that can catch the time and meet the needs by continuing the legal personhood discussions based on artificial intelligence without slowing down…


brozek, bartosz. “the troublesome ‘person’”, legal personhood: animals, artificial intelligence and the unborn (ed. kurki, visa a. j. – pietrzykowski, tomasz.), springer publishing, 2017, p.3-14.

european parliament. artificial intelligence and civil liability, pe 621.926, july 2020.

michalczak, rafal. “animals’ race against the machines”, legal personhood: animals, artificial intelligence and the unborn (ed. kurki, visa a. j. – pietrzykowski, tomasz.), springer publishing, 2017, p.91-103.

pietrzykowski, tomasz. “the idea of non-personal subjects of law”, legal personhood: animals, artificial intelligence and the unborn (ed. kurki, visa a. j. – pietrzykowski, tomasz.), springer publishing, 2017, p.49-68.

alicio erdogan