2018Murat Volkan DülgerArticle

Abstract

This work was written in Turkish. The summary on this page is a translation; the citation gives the original title in parentheses.

In this piece I will touch on what artificial intelligence (AI) is and where it came from, and, among the scenarios about the future of AI technology, I will discuss in particular its effects on the field of law. Of course, the subject touches us too. As is true of many of those reading this piece, the only work I know is practising as a lawyer and teaching at a university; in other words, I earn my living from professional legal services. In the light of these developments and debates, I therefore ask: are our profession and our livelihood slipping out of our hands?

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, Academic & Attorney

Today’s Artificial Intelligence Technology and the Reality of the ‘Robot Judge/Lawyer’:

Are We Losing Our Profession?

The latest point reached by today’s technology is artificial intelligence, the artificially intelligent entity. These days, scientific circles are discussing that artificial intelligences may be able to take part in mathematics competitions and perform surgical operations, that within ninety years they may even be able to carry out artificial intelligence research themselves, and that a hundred years from now these entities may even become like us. Can these claims be true, and could artificial intelligence really one day come to dominate the world? Or, as some think, are these possibilities too utopian? Yet had someone spoken years ago of the point reached today, would the technology we now live with and use ourselves not have seemed just as unbelievable? In order to give reasonable answers to all these questions, it is necessary to clarify what artificial intelligence is, how it developed, how we have arrived where we are today and, finally, what may happen in the future.

In this article I shall touch on what artificial intelligence is and where it came from, and then discuss, among the scenarios regarding the future of artificial intelligence technology, in particular its effects on the field of law. Of course, the subject of this article concerns us too. As is true of many of the people reading it, the only work I know is that of a lawyer and university lecturer; in other words, I earn my living from professional legal services.

In the light of these developments and debates, therefore, I ask the following:

Are we losing our profession and our livelihood?

In order to understand artificial intelligence, one must first understand the concepts of natural intelligence and of the brain, the source of natural intelligence. First of all, the source of mental activity is the organ called the ‘brain’, defined as the centre of sensation and consciousness and located in the central nervous system branch of the nervous system. What we call intelligence arises as a result of the complex processes carried out by the nerve cells in the brain. The word intelligence was first used in the work of Aristotle in the form “intelligence”, whose Turkish equivalent is “to understand”.

We can therefore say that the activity of understanding and comprehending lies at the basis of intelligence.

Although what intelligence describes and which brain activities it covers has been debated for many years, experts in the field share a common understanding that it is an ability that can be developed and that, owing to its connection with the brain, it is based on biological processes. Although there is no agreed definition, intelligence essentially denotes the abilities to grasp the relationship between concepts and objects, to think abstractly, to analyse, to adapt, to solve problems, to reason and to learn from experience. Accordingly, the stages of learning, analysing what has been learned and, at the end of that analysis, producing an output of some kind are signs of intelligence. If this is what is to be understood by natural intelligence, then the entity created by the technology called artificial intelligence is also expected to be able to carry out these processes.

WHAT IS ARTIFICIAL INTELLIGENCE AND HOW DID IT EMERGE?

Bringing inanimate objects together in various ways and turning them into intelligent, human-like beings has since very ancient times been one of humanity’s greatest curiosities and has occupied its mind. So much so that this curiosity even found a place in the mythologies of Ancient Greece and Egypt. However, because technology had not yet developed, this could at that time only be imagined. Much as, in fact, the possibility that entities with artificial intelligence might one day become lawyers, surgeons or writers occupies people’s minds today. Over time, the development of technology and computers made work on artificial intelligence possible. Indeed, the history of artificial intelligence can be seen to run parallel to the history of today’s computers, since computers are the first step in the development of artificial intelligence.

The most concrete step in this regard was taken by Alan Mathison Turing with the question “Can machines think?”. In his article entitled “Computing machinery and intelligence”, published after the Second World War, in 1950, in a British scientific journal called Mind, Turing discussed the possibility that machines could be intelligent and proposed a test for determining whether a machine is intelligent, the Turing Test, which is still valid today. Thus, for the first time, whether an inanimate entity can think became a subject of debate.

The concept of artificial intelligence, under its English name “Artificial Intelligence”, was first used, almost officially, on the proposal of John McCarthy at a conference held in 1956 at Dartmouth College in Hanover, New Hampshire. In this respect McCarthy is also regarded as the most important name in the history of artificial intelligence.

At the conference McCarthy spoke of the possibility of creating computers endowed with intelligence, and until the 1960s intensive work was carried out and, as a result of this work, important advances were achieved.

Marvin Minsky, who contributed to the formation of the field of artificial intelligence and developed the first artificial neural network computer, called SNARC, • took a rather optimistic approach in his book “AI: The Tumultuous Search for Artificial Intelligence”, stating that the problem of creating artificial intelligence would be solved within a generation.

On the other hand, although what happened in that period is now regarded as important for the birth of artificial intelligence, the failure of the work to produce any results led to a dark period for artificial intelligence. Despite the aims of developing software that could translate, solve problems and imitate human intelligence, no result capable of doing these things emerged. For the results aimed at were rather exaggerated for that period. The failure of the work to produce any results led both to a decline in hopes in this direction and to the cutting of the financial support allocated to research. Although work regained momentum after the 1970s, the real advances came after the 1990s. The defeat of the world-famous chess master Garry Kasparov by a computer developed by IBM, and a talking computer called “chatbot” fooling experts into believing it was human and thereby passing the Turing test proposed by Turing in 1950, greatly excited those interested in the subject.

What artificial intelligence, which has thus passed through a phase of birth and development to reach today’s technology, is and what it aims at can be summarised as follows: as we understand from all the work done since the 1950s, the basic aim is to create intelligent entities capable of imitating human intelligence. I stated above that the concept of intelligence denotes the abilities to grasp the relationship between concepts and objects, to think abstractly, to analyse, to adapt, to solve problems, to reason and to learn from experience.

Hence the basic and ultimate aim of work in the field of artificial intelligence is to create machines endowed with these

abilities. On the basis of the book by Russell and Norvig, definitions of artificial intelligence can be examined in four different groups: systems that think like humans, systems that act like humans, systems that think rationally and systems that act rationally.

In conclusion, artificial intelligence can be described as the performance by machines of the mental activities that are called “intelligence” when carried out by humans.

HOW DOES ARTIFICIAL INTELLIGENCE DIFFER

FROM THE INTELLIGENCE WE KNOW?

Machines created to imitate human intelligence need to be equipped with artificial intelligence algorithms. Does artificial intelligence, which tries to do what natural intelligence does, differ from it, or does it have aspects in which it is more advantageous?

It undoubtedly does. The first thing to be said in this respect is that artificial intelligence, unlike human intelligence, does not have a limited lifespan. As far as we know at present, since the human being is the only entity with an advanced natural intelligence, it can essentially be said that natural intelligence equals human intelligence. Since the human being is mortal, his intelligence likewise disappears with death. Artificial intelligence, by contrast, is everlasting as long as the system in which it exists does not change.

The fact that the information contained in a system with artificial intelligence can easily be copied and reproduced is one of the greatest advantages of artificial intelligence. While such copying is out of the question for human intelligence, the possibility of reproduction, dissemination to large audiences and hence easy availability that artificial intelligences offer will in future be the reason why these intelligences are largely preferred. The fact that artificial intelligence is built on a particular system ensures that it always reacts the same way to every situation. Human reaction, by contrast, is highly variable and irregular for various reasons, since a human being, because of his mental state or physical health, cannot always deliver the same performance. I think that this difference is important with regard to the possibility of using artificial intelligences for certain professional groups in the future. Instead of a teacher who cannot always teach at the same level, a surgeon who cannot operate for health reasons, or a judge who cannot make a sound decision because he quarrelled with his spouse that day, an entity equipped with artificial intelligence algorithms capable of making consistent decisions in every case may well be better: this is an undeniable fact.

Other differences that count among its advantages over natural intelligence are that artificial intelligence is objective, since it is not liable to be influenced by emotions, and that, since all its activity is systemic, the decisions it makes or the operations it performs can easily be documented by following the system.

Does natural intelligence, then, have no superior side at all? Of course it does. The best example of this is that human intelligence is creative. In order to obtain a particular result from artificial intelligence, information must be entered into the system. While the ability of artificial intelligences to produce a solution concerning information that has never been entered is rather limited, human intelligence, owing to its creative and generative nature, can reason about a problem it has never encountered before and produce a solution, whether right or wrong. The abilities of artificial intelligences to create and to solve problems, by contrast, are quite narrow in scope.

While human beings with natural intelligence are expected to use, and benefit from, the experience they acquire through perception in any problem they encounter, artificial intelligences remain limited to what is presented to them.

Thus, because artificial intelligence depends on the symbols entered and has been created with a particular system, although a number of advantages can be mentioned, there are also many differences owing to the fact that human beings have senses and possess a psychological and emotional side. Artificial intelligences can help us in many fields, can carry out certain activities in our place and may even, at times, do so better than human intelligence.

However, owing to their natural differences, I do not consider it likely that one will entirely replace the other in the near future.

ARE ARTIFICIAL INTELLIGENCES A RISK FOR HUMANITY? CAN THE DISASTER SCENARIOS COME TRUE?

Whether artificial intelligence is beneficial to us, or whether it harbours certain

risks for the future, is a matter of great debate. Alongside the scientists who describe today’s developments as a great success, there are also those who see them as the end of humanity.

The most striking statement on the subject came in an article written for the Independent newspaper by the renowned physicist Stephen Hawking. In his article, Hawking states that serious successes have recently been achieved in the field of artificial intelligence and that work is continuing, but says that caution is also needed in this respect. In that article he underlines that, if artificial intelligences continuously improve themselves and surpass human intelligence, this will pose a risk of the end of humanity, and that artificial intelligences must therefore be used in the way most beneficial to humanity.

Elon Musk, one of the most important names in the field of artificial intelligence today, holds a similar view. Musk, who has addressed the possibility that artificial intelligence will leave human beings behind, finds the only solution in human beings also acquiring an intelligence like that of artificial intelligences. Indeed, he has been working and investing to this end for many years.

Alongside scenarios such as artificial intelligence possibly being “the end of humanity”, there are also views that, on the contrary, describe artificial intelligence as a “helper to humanity”. Steve Wozniak, co-founder of Apple and one of the pioneers of personal computers, recently stated that artificial intelligence is very beneficial to human power and that there is no danger if we see this technology not as a risk but as a partner or friend.

Indeed, Ray Kurzweil, one of the pioneers of this field, also holds extremely optimistic views both on the development of artificial intelligence and its participation in daily life and on its possible effects.

The occurrence of such major developments in the field of artificial intelligence within such a • short period naturally also causes unease among people.

For the possibility that robot technology may be able to replace human beings entirely undoubtedly means, in some fields, the end of humanity.

Nevertheless, if it is used and developed in a manner beneficial to humanity, it is also possible, as Kurzweil and Wozniak foresee, for a force to emerge that is a partner or friend to humanity.

WHAT HAS ARTIFICIAL INTELLIGENCE BROUGHT US, IN WHICH FIELDS DO WE USE IT, AND WHERE MIGHT WE USE IT IN THE FUTURE?

Although the future of artificial intelligence is a much-discussed topic, it already has a highly developed structure today. Today we use artificial intelligences in many fields and technologies.

One of the common examples of the use of artificial intelligence is the ability to recognise and understand speech.

Converting human speech into digital format and understanding it by means of artificial intelligences is a very difficult and complex process.

Despite the difficulty of distinguishing differences in speech and, above all, of logically grasping what is said, this difficulty has today largely been overcome and success achieved. The most popular examples of speech recognition and understanding are Siri and Cortana. Thanks to artificial intelligence algorithms, these programs recognise your voice, largely understand what you say and can even respond to it.

There are also developments similar to the perceptual ability of the human eye.

Today artificial intelligences are used to see and detect dangerous situations in security camera footage and to respond to them.

There is no doubt that the greatest impact in this respect will be felt in the field of autonomous vehicles. Vehicles that can move without a driver thanks to their ability to perceive their surroundings must be able to process images and move

safely. Whereas self-driving vehicles were impossible until a short while ago, work in this area has today largely succeeded.

Natural language generation is another of the fields in which artificial intelligences are used.

The readability and comprehensibility of text generated from computer data have recently been greatly improved.

This technology is currently used mostly in customer services and business analysis. Thanks to various programs, the opinions of even millions of customers can be read via social networks, and what they want and what they are satisfied with can be learned.

In parallel with this, translation activities are also developing. Work continues on artificial intelligence algorithms that decipher the meaning of the sentence they read and translate it into another language.

It may be said that the most difficult field in which artificial intelligences are used is reasoning and decision management. For it is not easy for artificial intelligences to carry out, in all sorts of different situations, the act of decision-making, which is difficult even for human beings. Nevertheless, it is being said that in the future artificial intelligences will make more accurate decisions than human intelligence by calculating various possibilities. Recently, the “Dota 2” artificial intelligence developed by Elon Musk and the engineers of his organisation faced the best player in the world, and the professional player of the game suffered his first defeat against the artificial intelligence bot. This development is an important indication of artificial intelligence’s ability to learn and to reason by making its own decisions in the game.

Driverless vehicles, likewise, concern the ability of artificial intelligence to make its own decisions. Determining the direction in which the vehicle will turn, proceeding along the road it is to travel, or deciding whether to stop in an unexpected situation requires the power of reasoning. It is essentially from this point that the danger scenarios surrounding artificial intelligences arise.

Since it will no longer be possible to stop artificial

intelligence algorithms whose reasoning power has increased and which can make decisions by themselves, what will happen afterwards is a dark area. Moreover, the prospect that such a highly developed artificial intelligence will make decisions in place of human beings in daily life and replace them entirely, so that human beings will no longer be needed, is one of the most important reasons underlying the idea that the end of humanity may be brought about by artificial intelligences.

WILL ARTIFICIAL INTELLIGENCES

BE ABLE TO TAKE OUR PROFESSIONS

AWAY FROM US?

The fact that artificial intelligence has gained such a place in our daily lives and makes our lives easier also raises the question whether, over time, artificial intelligences will not merely help us but be able to do our jobs in our place. If the work being done to enable a self-driving vehicle to drive, and moreover more safely than human intelligence, succeeds, what will become of the profession of driver? If there is software that composes music especially for you, inspired by the songs you love, will there still be a need for musicians? What will translators do once there are artificial intelligence algorithms that can grasp what is said logically and simultaneously translate it into another language? If there are robots that can perform surgical operations more effectively than human beings, will surgeons be able to find work? Artificial intelligence technology is powerful enough to affect workers in every profession. For with the development of technology and the increase in data storage capacity, it does not seem at all impossible that artificial intelligences, whose skill levels and fields of skill are multiplying, will affect our professions to a great extent.

WHAT ARE THE EFFECTS OF ARTIFICIAL INTELLIGENCES ON THE FIELD OF LAW?

It is inconceivable that artificial intelligences, which will affect every branch of profession, should not affect a field such as law. The most important problem is how entities equipped with artificial intelligence algorithms are to be regulated by law. How should these entities, which are not human but cannot be regarded as things either, be classified and regulated? Will it be possible to speak of fundamental rights of these entities, as of fundamental human rights? Will robot rights and freedoms become a field of law that needs to be regulated in the future?

To which law will offences committed by robots be subject? If Sophia, the human-like robot produced by the Hong Kong company Hanson Robotics and recently declared by the Saudi Kingdom to be the first robot to receive citizenship, has a family in the future, as she dreams of doing, under which civil code will the robot Sophia found a family?

As can be seen, dozens of questions come to mind concerning the most fundamental fields of law, from constitutional law to criminal law and civil law. Although there are no clear answers to these questions today, it is an inescapable fact that answers will have to be found in the future. In the face of such rapidly developing technology, innovations must also be made in the field of law, and law, the most important component of social order, must also begin to work in this area.

CAN ROBOTS WITH

ARTIFICIAL INTELLIGENCE BECOME JUDGES

AND LAWYERS?

With an interesting recent development, we were all suddenly confronted with this question. Among the possibilities being considered is that the legal profession will disappear in the near future and that robots will be able to make better decisions than judges.

Indeed, an important development in this direction has already taken place. The law firm Baker & Hostetler in the USA, which employs some fifty lawyers, announced in a statement that, by employing IBM’s artificial intelligence software ‘AI Ross’ in the field of bankruptcy law, it had become the first in the world to hire a robot with artificial intelligence. It is also claimed that other law firms have concluded contracts with IBM for Ross.

In producing Ross, IBM used a technology it had developed itself, called Watson, and equipped the robot Ross with capabilities such as reading, understanding and producing solutions to problems.

It is said that Ross, not limited to these, will learn more over time with experience and that its learning process will never end. With this development, it appears that the situation is serious. The artificially intelligent entity is entering our lives more and more with each passing day.

Another study on the subject is being carried out on an artificial intelligence system developed by scientists from University College London and the universities of Sheffield and Pennsylvania. As part of the research, English-language data sets of 584 cases brought on the ground of violations of Article 3 of the ECHR on the prohibition of torture, Article 6 on the right to a fair trial and Article 8 on respect for private life are being entered into this system.

Depending on the algorithms created, the system delivers a ‘violation’ or ‘no violation’ decision for the cases. The reason for choosing these articles is that they concern fundamental rights and that a great deal of data is available on them. With its decisions, the system under study was able to predict correctly, at a rate of 79 per cent, how cases before the ECtHR would be decided.

The scientists who carried out the research stated that, although the result is exciting, they do not think that lawyers or judges will be replaced in the near future and that they will continue their work, adding that this work could nevertheless help in deciding cases.

The rate at which the system under study predicts the Court’s decisions is quite astonishing. For although the possibility that there will be no further need for lawyers, judges and courts may for a moment come to mind, the factor that human intelligence has an emotional and creative side comes into play here too. It should not be forgotten that ‘discretion’, which occupies an important place

in the field of law, cannot be exercised by robot judges and may be very limited for the present. For this reason, it is unlikely that there will be no need for the professions of lawyer and judge in the near future. Nevertheless, the use of the systems developed in the field of law as well, their ability to assist judges, and the use of artificial intelligences at the points where they are superior to natural intelligence should undoubtedly be regarded as a development beneficial to humanity.

With technology developing so rapidly, artificial intelligence, whose history arose in parallel with the history of the computer and which came onto the agenda with the question whether the abilities of human intelligence can be performed by machines, and the entities possessing such intelligence, have come to occupy a large place in today’s technology. While they are used in many fields and branches of profession, scientists are rapidly continuing their work so that they will be used even more. Today artificial intelligence has become one of the fields attracting the most investment.

That is why I think that work in this field should be carried out more intensively in Türkiye. In addition, scientific studies should be conducted on the possible effects of artificial intelligences on the professions. In particular, I believe that, as regards the place of artificial intelligences in the field of law, regulations should be introduced following comprehensive studies, and that these entities may be beneficial in developing and improving the law.

Assoc. Prof. Dr. Murat Volkan DÜLGER Academic & Attorney volkan.dulger@dulger.av.tr

Download PDF (in Turkish)

More publications on this topic