Introduction
This work was written in Turkish. The summary on this page is a translation; the citation gives the original title in parentheses.
In parallel with developments in technology, great and significant progress is also being made in the field of artificial intelligence (AI) and AI entities. Indeed, AI algorithms are renewed and improved every day, experiments are carried out on the subject, and many of them succeed. The main reason for this work is the wish to use AI in many quite different fields, above all IT, the internet and electronics, because of the advantages I explain below. Moreover, even at present AI algorithms are seen to be used intensively in the conduct of economic activities. Developed countries are known to be investing large sums in this technology today, and this fact alone is in essence enough to show the importance of AI technology.
As AI has entered our lives to this extent, curiosity about the subject has grown in equal measure. One of the questions that arouses most curiosity is how AI will affect the field of law. This is because regulating by legal instruments the AI technology currently in use, and that which is planned to be attained in the future, is a necessary and compelling need. In this piece I will try to discuss the need for legislation to address the legal status and personality of AI technology, and in particular of AI robots and of entities that will have the capacity to reason; their civil or criminal liability where they are involved in an unlawful act or an offence; and whether there can be rules that apply among themselves.
Full text
This text is a translation of the Turkish original, prepared for this website. Numbers in square brackets are the page numbers of the original; sources in the footnotes are given as in the original. For citation, the original publication (PDF) is authoritative.
Assoc. Prof. Dr. Murat Volkan Dülger*
In parallel with developments in technology, great and significant progress is also being made in the field of artificial intelligence and artificially intelligent entities. Indeed, artificial intelligence algorithms are renewed and improved every day, experiments are carried out in this field, and many of those experiments succeed. The main reason for this work is the wish to use artificial intelligence, on account of the advantages I shall explain below, in many different fields, above all information technology, the internet and electronics. Moreover, even as matters stand, artificial intelligence algorithms are seen to be used intensively in the conduct of economic activities. It is known that developed countries today invest large sums in this technology, and this fact alone suffices to demonstrate the importance of artificial intelligence technology.
With artificial intelligence having entered our lives to such an extent, curiosity about the subject has grown in the same measure. One of the greatest questions is how artificial intelligence will affect the field of law. For the regulation, by legal instruments, of the artificial intelligence technology that is currently in use and that is planned to be attained in the future is a necessary and compelling need.
In this article I shall try to discuss the need for statutory regulation of the following matters: the legal status and personality of artificial intelligence technology, and in particular of artificially intelligent robots and of entities that will possess the capacity to reason; their civil or criminal liability where they are involved in an unlawful act or an offence; and whether there can be rules that apply among themselves. I shall set out my views on which criteria should be taken as the basis when such regulation is made and on how and in what form it is to be carried out. Finally, I shall mention whether any regulation on this subject already exists.
The approach of the legal order to the privacy and confidentiality gaps that artificial intelligence technology may create
and the regulation that needs to be made
Ever since it came into existence, humankind has striven to make tools that would meet its needs, make its life easier and help it. Making these tools requires knowledge, skill and experience. This is the point at which the concept of technology comes in. In short, technology is the name given to the means which people need in order to make, produce or develop the tools and implements that will help them, and which develop and multiply as science advances. Through research, studies and experiments, technology develops day by day and reaches astonishing dimensions.
Thanks to scientific progress and technological developments, the advances in the world of information and communication have seriously changed our lives and made them easier. While computers, telephones and the internet make it possible to communicate with the other end of the world, means of transport have made every corner of the world accessible. In the field of health, activities such as the operations performed, X-rays and scans, blood tests and DNA research are wholly dependent on scientific progress and technology. It can thus be said that technology has made great contributions in terms of diagnosis, methods of treatment and the speed of access to treatment, and therefore in terms of the lengthening of human life. In the field of law, and of criminal law in particular, it has become almost impossible to conduct proceedings and to catch offenders without using technology. Technologies such as security cameras, blood traces and fingerprints, and tracking systems are frequently used in criminal investigations and prosecutions.
As can be seen, technology and what it brings have become indispensable in communication, transport, health, law and many other fields. Indeed, it is a plain fact that technological products, with the conveniences they provide, have become part of our lives. So much so that a life without the benefits and conveniences provided by technology is no longer conceivable. But is the technology that makes our lives so much easier really that innocent; does it do no harm at all? Unfortunately that is not the case. Technology takes away as much as it brings, and some even claim that it takes away more. Although I do not share that view, it cannot be denied that alongside its benefits it also does a good deal of harm.
Weapons of war that have developed along with technology, air and environmental problems, the depletion of underground and surface resources, and the harm done to living beings by the electromagnetic waves emitted by technological devices are the first of the difficulties that have entered our lives with technology to come to mind. Besides these concrete and material examples, the greatest harm in fact occurs in people's social lives. It is observed that, following the conveniences provided by technology, people have become lazier and spend more time in the digital sphere. This has brought with it many problems such as depression, stress and insomnia. The human being who, thanks to social media, can express himself and share his thoughts more easily, more quickly and with wider audiences has, on the other hand, turned into a virtual and dependent being who does not communicate face to face. As a result, a new kind of illness has entered our lives under the name of technology addiction and has begun to be regarded as the illness of the age. Many organisations have taken action and begun work with the aim of combating technology addiction.
The most important point at which the benefits and harms of technology meet, however, is the exchange of information. With the development of information and communication technologies, especially by way of the internet, the exchange of information has become very easy and fast. Today it is possible to store, process and transfer very large quantities of data within seconds. At the same time, the person to whom data relate can easily be reached by starting from the data. Alongside the benefits it provides, this situation has raised the danger that information belonging to individuals may fall into the hands of third parties who have no connection with or interest in that information and may be used by those third parties, lawfully or unlawfully, and concern about this has begun to be felt. Although these concerns originated with simpler technologies, it is with today's technology that they have grown and reached serious proportions.
We see that the latest point we have reached in technology today consists of developments that, only a short while ago, were hard even to imagine and, even if imagined, were regarded as impossible to realise. The point reached in the field of artificial intelligence in particular is both exciting and, in the face of possible scenarios, thought-provoking. For it is said that artificially intelligent entities will be able to replace many professions, to perform surgical operations and, within ninety years, even to carry out artificial intelligence research themselves, and indeed that in a hundred years these entities may be like us. What is expected of artificial intelligence is that it should be able to carry out the stages of learning, analysing what it has learnt, and reaching a decision by interpreting the output of that analysis and, where an artificially intelligent robot is concerned, even of acting. When one considers the artificial intelligence work continuing for these purposes and the existing products of that work, these possibilities do not seem all that impossible.
There are those who argue that the foreseen future of artificially intelligent entities will, just like developments in technology, be beneficial and helpful to humanity; there are equally those who, quite to the contrary, regard it as dangerous for human beings. What, then, is the source of this danger? What risks could there be in artificially intelligent entities that will be able to carry out research like human beings, to assist people in the professional activities they perform, and to provide many conveniences in a person's life, such as being able to predict even the time at which he will drink his coffee and when he will go where? In fact, the answer to this question lies precisely in the question itself. Although the things I have listed make one think at first sight how wonderful artificial intelligence technology is, they in fact also describe the possible dangers. How can an artificially intelligent entity know the time at which you will drink your coffee? How is it possible for it to remind you of the medicine you have to take and the time at which to take it? How can it know your family and your close circle? All these questions have a single answer: data, and in particular personal data! Artificial intelligence technology performs mental operations on the data loaded into it and produces results. It therefore essentially starts out from data, and here we encounter the concept of big data. The realisation of artificial intelligence depends on the development of artificial neural networks, and this in turn depends on the generation, storage, accessibility and processing of large quantities of data.
Artificial intelligences, which engage in activities that imitate human intelligence by using data, can in many different respects perform these activities better than human intelligence. The respects in which artificial intelligence is superior to human intelligence in the process of analysing the data loaded into it are these: unlike natural intelligence, that is, human intelligence, it has no fixed lifespan or capacity; the data in a machine equipped with artificial intelligence algorithms can easily be copied; and, being built on a particular system, it can give the same, objective responses in every situation.
It is feared that artificial intelligences, which it is foreseen will be able to use data belonging to individuals so effectively and without restraint, may, as a result of the operations they carry out on those data, give rise to unlawful uses and to human rights violations. Indeed, the fact that human rights are already being intensively violated and put at risk in the fields in which artificial intelligence algorithms are used reinforces these fears. The point at which the concerns are chiefly felt and concentrated relates to the sphere of privacy and secrecy, within the framework of the confidentiality of individuals' private lives and the right to respect for private life.
The right to privacy was defined by the 1970 “Declaration on Mass Communication Media and Human Rights” as the right to live one's own life with a minimum of interference. A person's private life, physical and mental integrity, honour and reputation, and family life fall within its scope. The case law of the ECtHR also states clearly that the right to privacy is likewise protected by Article 8 of the European Convention on Human Rights, which bears the heading of the right to private life and family life.
Accordingly, a person should be able to lead his life with the least possible interference in his own integrity and in the spheres of his private and family life. Yet today's technology and the future of artificially intelligent entities are being produced and designed in a way that is capable of interfering with this right to the maximum degree. The reason is that every time we use the technologies in question we leave digital traces behind us. Every step we take in the digital sphere and, in connection with it, on the internet – information relating to our most intimate sphere, such as which websites we visit and what we search for, which products we look at on the shopping sites we visit, and even where we eat and whom we are with – is constantly being recorded. The curious thing is that, more often than not, we ourselves leave lying about, with our own hands, this information which we would not want anyone to know or to reach.
Today the problems concerning the protection of personal data have come to be better understood because of the dangers that have arisen largely with the development of technology, and, as a result of the European Union's work in this direction, particularly from the 1980s onwards, the right to the protection of personal data has taken shape. In our country too, personal data are protected by Article 20 of the Constitution, following the amendment of 2010, and by Law No. 6698, which entered into force on 7 April 2016, and the related regulations. In addition, the unlawful recording and dissemination of personal data has been regulated as an offence by the Turkish Criminal Code since 1 June 2005.
Nevertheless, there is no doubt that these protections will prove inadequate as technology develops and artificially intelligent entities enter our lives in earnest. I think that in the near future we shall be confronted with the unlawful use of our personal data by artificial intelligences. Considering the point that has been reached, I foresee that the possible dangers I mentioned above are not very far off. For this reason I think that work on technology and artificial intelligence must also begin at the level of law, that ethical principles must be determined, and that regulation must be undertaken.
In this context, the debate on the legal status of artificially intelligent entities must be resolved first. Next, since it is no longer doubted that in today's world life cannot go on without the use of personal data, it is certain that artificially intelligent entities too will need personal data. The cases in which personal data may be recorded, processed and/or transferred by artificial intelligences must therefore be determined by legal provisions. For example, the conditions under which an artificially intelligent entity may transfer personal data in its possession to another artificially intelligent entity must be foreseeable. After that, the civil and criminal liability of these entities must be regulated for cases in which they violate fundamental human rights, above all privacy and the right to respect for private life.
Criminal liability of the artificially intelligent robot (is it possible?)
I must first state that there are no statutory provisions determining the legal status and the civil and criminal liability of artificially intelligent entities, but that the need for regulation will arise as technology develops. The reason is that our legal order recognises only the distinction between persons and things that has come down from Roman law. Yet just as artificially intelligent entities cannot be made subject to the rules to which human beings are subject, neither can they possibly be assessed within the definition of a thing. There is no doubt that artificially intelligent robots go far beyond the definition of a simple thing. For this reason, the limits of the traditional understanding of law must now be transcended and work must begin on rules determining the rights and obligations of artificially intelligent entities. However, since no such regulation yet exists, at this point we can speak only of what ought to be.
In order to answer the questions whether artificially intelligent robots can bear criminal liability and, if so, how this is to be determined, attention must first be drawn to the distinction between autonomous and semi-autonomous machines on the one hand and artificially intelligent entities in the true sense on the other. Solving this problem is simple as far as autonomous and semi-autonomous machines are concerned. Since these machines have no consciousness and are controlled by natural persons, the natural persons who control them or, as the case may be, who produce them will be liable for the damage they cause. For these machines act according to the commands given by natural persons, and so one cannot speak of any civil or criminal liability on their part.
The real problem in determining liability is what is to happen in the case of artificially intelligent entities in the true sense. If an artificially intelligent robot capable of taking the final decision itself is the driver of a car and, in order not to crash the car, hits someone else, who will be the perpetrator of the offence of intentional or negligent killing? In another scenario, if a robot capable of performing surgical operations causes the death of a patient through its neglect, who will be liable? The artificially intelligent robot itself, its owner, if it is accepted that it has one, or the manufacturer of the robot?
If it is accepted that the artificially intelligent robot has an owner, the robot has to be treated as having the status of a thing. Yet these artificial intelligences have no controller and no one who gives them commands. They are therefore able to take their own decisions. Consequently, one cannot speak of their having an owner. Nor should it be possible to hold the manufacturer liable for damage arising from conduct that the robot has carried out by its own conscious acts. To regard a manufacturer who is in no way culpable as the perpetrator of the offence is contrary to the principle that offences and penalties are personal. Therefore, an artificially intelligent robot that acts consciously of its own will and, as a result of that act, brings about an offence fulfilling a statutory definition must be regarded as the perpetrator of that offence. In our present system of criminal law, however, only natural persons can be the perpetrator of an offence.
As can be seen, when we assess the different possibilities in determining liability, the statutory provisions currently in place are unable to answer any of them and reach an impasse at some point. For this reason, artificially intelligent robots must also be defined in the legal order alongside the distinction between persons and things, and their legal status must be determined. The point to be borne in mind here is that, rather than a general definition, they need to be distinguished by grading them, for instance as autonomous, semi-autonomous and truly artificially intelligent entities. For the degree of artificial consciousness and of the capacity to decide for oneself is not the same in all of them. The extent of liability may therefore differ as well.
In the light of this information, where an artificially intelligent robot has, for example, caused a fatal accident, liability should be determined on the basis of the following stages:
• First, the definition and legal status, as determined by the Constitution and the Civil Code, of the robot that caused the accident must be established. Did the accident occur as a result of a command that was given, or can one speak of the robot's own consciousness and will? In the most extreme case, are biological human factors present? These possibilities must be assessed.
• If it is concluded that the accident occurred as a result of the conscious acts of the artificially intelligent entity, it must be examined whether the elements of the offence are present.
• Finally, the criminal sanction determined in accordance with the legal principles under which the liability of the artificially intelligent robot is provided for must be applied to the offence committed in the specific case.
Another problem that needs to be discussed here is how effective the criminal sanctions laid down for human beings will be on robots. The essential starting point of criminal sanctions is their deterrent effect before the stage at which the offence is committed. A human being who thinks that he will be sentenced to imprisonment may desist at the thought of being separated from his family and loved ones; but can this have a deterrent effect on an artificially intelligent robot? Or will there, further in the future, be robots that have feelings just like a human being?
After the stage at which the offence has been committed, the aim of criminal law is directed towards reforming the offender. The aim is that the offender sentenced to imprisonment should, during that time, come to regret the offence he committed and grasp that he must not offend again. Will it be possible to speak of the remorse and reform of an artificially intelligent robot?
Where a security measure is imposed together with or separately from a penalty, what meaning and what necessity does it have for a robot? When every problem can be solved by a software update, how realistic and sensible is it to leave an artificially intelligent robot to rot in a corner as a security measure?
As can be seen, just as the legal status and the civil and criminal liability of artificially intelligent entities need to be provided for by statute, how this is to be done also poses a major problem. The laws to be enacted must take account of the nature and essence of artificially intelligent entities and be capable of meeting their needs.
Whether a difference arises in criminal law between damaging ordinary objects and damaging an artificially intelligent robot, whether or not it has human-like characteristics:
the problem
I shall approach this problem by first taking the concept of a thing as the basis for the expression ordinary objects. Where objects that can be characterised as things are damaged, the offence of “damage to property” regulated in the Turkish Criminal Code arises as a result of that damage. However, where what is at issue is an artificially intelligent robot that can be regarded as falling within the scope of things, these cases too have to be resolved by the provisions specifically laid down for artificially intelligent entities. It can nevertheless be said that this should be regulated in the form of an offence similar to the offence of damage to property.
On the other hand, there may be artificially intelligent entities that are not human-like but that mean far more to a person than a mere thing. How correct is it to characterise merely as a thing a vehicle endowed with artificial intelligence that has a remote communication capability, that will pick you up wherever you wish and choose the most comfortable route free of traffic, that knows the coffee you like and can collect it from a café on the way before you arrive at home or at the office, and into which you can therefore code your personal preferences? For in that case you will not merely have damaged a person's car; you will have inflicted on that person a far greater harm than in the case where you had damaged his car.
A robot with human-like characteristics, by contrast, should be understood as an artificially intelligent entity that can perceive like a human being, that learns, that can decide for itself and that acts consciously. Will the entity itself be able to be the direct victim here? Will the perpetrator be punished for damaging property or for harming a person? If, where a robot possessing consciousness and the capacity to reason is harmed, the robot itself can be assumed to be the one directly harmed by the offence, it must be said that an offence similar to the offence of “injury” provided for in the Turkish Criminal Code will arise. The perpetrator must also be punished according to the legal status provided for an artificially intelligent entity with human-like characteristics.
It is apparent from these possibilities that three different types of artificially intelligent entity are envisaged: the robot that can be characterised as a thing, the one that has no human-like characteristics but represents more than a simple thing, and the one that has human-like characteristics. There is no doubt that these types will multiply as technology and artificial intelligence develop. The problem, however, is that all artificially intelligent robots are assessed under the same roof, without distinction, by means of the expression whether or not it has human-like characteristics. Yet such a regulation would yield nothing more than a single step beyond the present order. Besides, the very core of the rules ought to be whether or not the artificially intelligent entity has human-like characteristics. The real problem and impasse arise at this point. Consequently, rather than being gathered under one general heading, the principles and statutory provisions to be created in relation to artificially intelligent entities need to be comprehensive in nature and take a broad perspective.
Could criminal law norms applicable among robots themselves
be laid down in the future?
For the reasons I have stated above, the existing legal order is unable to respond to any dispute that may arise among robots themselves. Faced with a scenario such as an artificially intelligent robot injuring another robot, or another robot damaging property, it is not possible to respond with the provisions currently in place. For this reason, artificially intelligent robots too need to be defined as subjects of law, and the fate of the disputes that will arise among them and how those disputes are to be resolved need to be determined.
At the same time, I do not think it right to envisage a system that applies solely among artificial intelligences and regulates only the rights and obligations of the classified artificial intelligences towards one another. For such an approach and assumption would be at odds with the theory of the offence, and a dual legal order would emerge. Instead, once artificially intelligent entities have been defined and their legal status determined, it would be appropriate to adopt criminal provisions consistent with the general principles of law to which these entities will be subject.
Is artificial intelligence a danger or an advantage for cybersecurity?
First of all, as I explained in detail above, it must be accepted that technology and technological products in general pursue the aim of making our lives easier, but that, while making our lives easier in many different respects, they also bring many dangers with them. The same holds for artificially intelligent entities. The artificially intelligent entity that will make our work easier by assisting us in the professional activities we carry out also harbours a great danger for our future, in that at the next step, once it possesses and can perform all the qualities the professional activity requires, it may take our jobs away from us. Artificial intelligence algorithms, which now enable us to carry out in seconds operations that years ago took up our days and hours, are today also capable of becoming the subject of possible human rights violations by recording our personal data in very large volumes within a very short time. Artificial intelligence technology, which is used in every area of the IT sector and contributes to that sector's development, will on the other hand be capable of being used as a tool of cyberattack.
We therefore see that, in every respect, artificially intelligent entities harbour certain dangers alongside their very important benefits. But it would be a very great mistake to turn our backs on technology on account of these dangers. Instead of being against technology, one must be able to keep up with it and turn it into an opportunity. Indeed, keeping up with technology has today become a necessity. Countries and societies that do not use technology and do not avail themselves of its blessings cannot develop or keep pace with the new world. The best example of this is that the countries which were unable to keep up with the technology that came about in Europe at the end of the 18th century with the use of steam power in industry and with mechanisation have still not managed to cease being third-world countries.
For the reasons I have explained, however many dangers technology may harbour, it is not something to be feared and fled from but a necessity that must be followed. In order to derive the greatest possible benefit from technology, the dangers in question must also be foreseen in advance as far as possible and prevented under the State's positive obligation, and the rights of individuals must be protected. At a time when no specific field, including daily life, can continue or function without technology, running away from technology because of the risk factors it harbours is not a realistic approach; it is also tantamount to trying to reverse the current.
Legal regulation policy on artificial intelligence technology in the Turkish legal system:
what should it be?
I noted above that artificially intelligent entities need to be defined at the statutory level and their rights and obligations determined, and moreover that this regulation needs to be graduated and comprehensive, taking account of their different types. My view as to which legal principles should govern this regulation and what form it should take is as follows, in order of priority:
• First, the subject needs to be addressed in the context of constitutional law. Here the legal status of artificially intelligent entities must be defined and their fundamental rights and freedoms clearly determined.
• Next, civil law, which is the foundation of law, must come into play. A heading along the lines of a “law of artificially intelligent entities”, similar to the law of persons, should be provided for, and under this heading the provisions concerning the beginning and end of the existence of artificially intelligent entities, family law, the law of succession and the law of property should be laid down.
• I am of the opinion that criminal law should follow next in this order. It must be made clear what is to be done in a criminal law dispute involving artificially intelligent entities, and how and by reference to what their liability is to be established.
• Finally, since artificially intelligent entities will have social, economic and cultural relations and activities, regulation is also needed in other branches of law such as commercial law, labour law and the law of obligations. Indeed, when one thinks of artificially intelligent robots that acquire citizenship, it can be said that even international law has to come into play.
Footnotes
- Academic/Attorney-at-law ↑
More publications on this topic
Dülger, Murat Volkan / Çetin, Selin / Aydınlı, Celal Duruhan, “Algorithmic Decision-Making and Data Protection” (Algoritmik Karar Verme ve Veri Koruması), 2020.
Dülger, Murat Volkan, “Artificial Intelligence Technologies and Data Protection Law” (Yapay Zeka Teknolojileri ve Veri Koruma Hukuku), Academia.edu, 29 September 2019.
Dülger, Murat Volkan, “Emotional and Sexual Relations Between AI Entities and Humans: Are Sex Robots Replacing People?” (Yapay Zekalı Varlıklar ve İnsanlar Arasında Duygusal / Cinsel Yakınlaşmalar: İnsanların Yerini Seks Robotları Mı Alıyor?), Hukuk ve Daha Fazlası Dergisi (Law and More), No. 6, June 2018, pp. 64–70.
Dülger, Murat Volkan, “Today's Artificial Intelligence Technology and the Reality of the ‘Robot Judge / Lawyer’” (Günümüz Yapay Zeka Teknolojisi ve 'Robot Yargıç/Avukat' Gerçeği), Hukuk ve Daha Fazlası Dergisi (Law and More), No. 5, January–March 2018, pp. 4–9.
Dülger, Murat Volkan, “A Law-Fiction Essay: The Legal Liability of an AI Entity (Could It Be?)” (Bir Hukuk-Kurgu Denemesi: Yapay Zekalı Varlığın Hukuki Sorumluluğu (Olabilir Mi?)), Hukuk ve Daha Fazlası Dergisi (Law and More), No. 2, July–August 2017, pp. 4–10.
Dülger, Murat Volkan, “The Use of Three Technologies in Money Laundering: Fintech, Crypto Assets and Artificial Intelligence” (Suç Gelirlerinin Aklanmasında Üç Teknolojinin Kullanılması: Fintek, Kripto Varlıklar ve Yapay Zeka), conference presentation, 2026.

