Scale Automation in the Processing and Storing of Legislation and Case Law

Introduction:

Currently, legislation and case law are written in human language. However, it could be possible to (re)write it in a way that makes it understandable and usable for computers. The possibilities that this would create are vast. It would allow for large scale automation in the processing and storing of legislation and case law. We will further explore the possibilities later in this paper. We will examine the positive and negative aspects of binary translation within the legal system, look at possible uses for it as well as the potential problems that come with it.

Firstly, to get a clear view of the difficulties that arise with the binary translation of legislation and case law it has to be clear what binary translation is about. It is sensible to set out what a binary code is asswel. After we have explained binary code and binary translation we will try to explore potential reasons of why justice systems around the world are turning towards digitisation. We will discuss the general difficulties that arise for the judiciary and for the tax authorities when they use automation and digitization. Furthermore, we shall examine the opportunities binary translation creates, combined with the problems that might arise simultaneously. Hereafter we try to project these problems and difficulties on the translation of legislation and case law in binary code.

Before we get in to the precise meanings of binary code and binary translation it might be useful to outline some other definitions that will be used in this paper. First of all, when we talk about artificial intelligence (AI) we assume the definition as it is given in the Oxford Dictionary: “The capacity of computers or other machines to exhibit or simulate intelligent behaviour; the field of study concerned with this.”

The binary system is a numeral system that consist of only two numbers instead of the ten numbers in the decimal system that is used in the day to day of life of people. The meaning of the word “binary language” gives a good representation of what it actually is, namely a language that has two options, being “0” and “1”. It has its origin in latin language where “binarius” means “consisting of two”. A binary code consists of 0’s and 1’s and every number in the code is called a bit. We use the binary system for coding computers because of the way computer memory cells work. These cells are only able to turn off and on, this is what the ‘0’ and ‘1’ represent. Even the newest and most advanced programming languages are in their cores based on the binary system. Computer programming languages are what allow us, as human beings, to instruct a computer on what we want it to do for us. Similar to human languages, there is a large amount of programming languages that exist, each with its own uses and possibilities. However, the part of the language that a computer understands is “binary”.

The main reason why we do not write code in binary is because of simplicity. Humans do not want to spend hours upon hours writing 0’s and 1’s in order for a computer to do what we want it to do. This is why we have invented so called high-level programming languages. High-level programming languages hide many details of computer operations. Whatever we write in a piece of code written in a high-level programming language, for example Python or C#, gets translated into binary in order for the computer to understand it. This is done by a compiler. A compiler is a program that takes the (high-level) code written by a human and converts it in such a way that the computer can understand.

Binary Translation

Now that it is outlined what a binary code is, the concept of binary translation has to be explained. Binary translation is the translation of instructions from a certain source that are translated into another instruction set. In the case of this paper, binary translation would be the process of turning case law and legislation written in human language in to binary language for a computer to understand.

Why are justice systems around the world turning digital?

The global society is currently experiencing an enormous shift towards digitisation and the world’s courts are no exception to this. Today’s courts are moving towards a concept of digital justice to try to deal with the large amount of problems they are facing. First and foremost, courts around the world are being pushed to their limits by the amount of cases they have to handle with fewer employees. This inevitably results in a giant backlog of cases. One example of this can be seen in the US, where the immigration court system has seen its workload increase by 146 percent in the last 10 years. As a result of this increase, in July 2015 the average waiting time per case was 627 days. This is especially critical when it comes to criminal law where the accused may end up spending years in detention waiting for a verdict.

Another problem that arises is the overuse of paper. Traditionally, every court is paper based which means that an increase in workload includes an increase in paper used. Saving, sharing and transporting physical documents is a lot more expensive than doing this with digital files. Furthermore, a heavy reliance on paper can hinder the execution of justice since a single missing page in a bundle of documents can result in a case being adjourned for lack of evidence.

Digitizing the legal system results in higher efficiency and lower costs in the long run. This can not only be seen in the courthouses themselves but in every step that is taken before a case gets to court. There are thousands of paralegals and other employees assisting lawyers by sifting through hundreds of pages of case law and legislature, looking for that one rule or quote that gives them the edge in court. This is of course extremely costly and time consuming. There are already tools being developed that can be used in order to speed up this process, such as Deloitte’s “TAX-I” system. According to the website of Deloitte “TAX-I uses artificial intelligence (AI) to analyse thousands of tax cases of the European Court of Justice, relate them to similar cases, summarize them and even predict how a court would rule in a case”. A system such as TAX-I is only possible if it has sufficient data to work with. Currently, the data has to be manually collected and translated in such a way the AI is able to comprehend and utilize it. However, this process could be simplified if legislation and case law would be immediately understood by a computer (or AI system). This is where the binary translation of legislation and case law would make a very big, and very positive impact.

Digitization and automation of legal procedures and court decision making. The jurisprudence is also affected by an increased degree of digitization and automation. In the Netherlands K.E.I. (Kwaliteit en innovatie, which means quality and innovation) was introduced. The thought behind K.E.I. is that legal procedures should be more efficient. This could lead to the saving of costs for the courts. Furthermore, the digitization of legal procedures gives extra opportunities for citizens as well. It will be easier to add extra documents to the procedural record at any given time. Moreover, it will be possible to add digital documents to the annexes as well, such as videos and audio files. However the process of digitization of legal procedures is one with setbacks as well. The K.E.I. for instance does not function optimally according to van Mierlo.

Automation of the procedural law can improve the judgement of a judge too by making the process more standardized. Using a standardized procedure can lead to a more uniform jurisprudence, which can lead to more legal certainty. However this improvement of efficiency and uniformity can come at a cost. According to Feteris the use of technology should mostly be one of a supporting nature in forming a judgment by a judge. A judge should take all the relevant facts into account by forming a judgement, he should not be limited by a standardised decision making process. If necessary a judge should have the possibility to form his judgement in a way he thinks is right.

Artificial intelligence can play an important part in the jurisprudence as well, according to Van Oostrom-Streep. Since the law is a logical framework of rules. Court decisions should also be based on a logical conclusion.

The Tax authorities and digitization and automation of operating systems. Digitization and automation play an important role for the Tax Authorities as well. Because of the fact that the Tax Authorities have to deal with vast amounts of data they had to create a largely automated working environment to be able to cope with this amount of work. The Dutch Tax Authorities use fully automated decision-making rules. This algorithm makes decisions without any human interference. This is because a computer is better at math and can also work faster and more efficient than a human can. The Tax Authorities use artificial intelligence (AI) to make the decisions. The computer system is instructed by humans nonetheless. It is instructed to make decisions using reason and interpretation, however it is not self learning. As a result of the fact that the AI is not self learning, in theory it is still possible to reconstruct the decision making process. The Dutch Tax Authorities mainly use AI to make a risk profile to determine which taxpayers have to be audited. This could raise questions from the taxpayer, it is likely he will wonder why he is being audited and someone else not. According to van Eck it would be better if the Tax Authorities clarify to the taxpayer what the basis of the audit is.

The digitization and automation comes with a few other complications for the taxpayer as well. For example, tax payers are not able to verify how a certain decision is made. According to van Eck the legal protection falls short because of this. The Tax Authorities do not explain the decision making process, because of this it is not possible to confirm if the decision that is made is in line with the law.

Less self-sufficient taxpayers face other difficulties. Since sometimes they do not fully understand the digital tax forms and are not able to use digital communication. The automation of the work of the tax authorities is necessary but cannot come at the expense of the legal protection of the taxpayer, according to Stevens. These automated processes lead to a less personal treatment of every case. Personal circumstances of the taxpayer are not taken into account because of these rigid operating processes. This is not preferable because it can lead to less legal protection for the taxpayer.

Possibilities and limitations of the binary translation case law and legislation As stated before, legal procedures are getting more and more digitized. Because of this there lies an opportunity to make the legal procedures even more efficient by using a language that is readable for computers as well. By translating legislation and case law into the binary language it might be possible to use the full potential of computers and the advantages that come with it. First we will discuss the possible advantages of the binary translation of case law and legislation. We will discuss legal equality, legal certainty and budgetary elements. Thereafter we will discuss the limitations that come with the binary translation of case law and legislation.

Legal Equality

Legal equality is one of the fundamental principles on which most western legal systems are based. In short, legal equality is the principle that each independent individual must be treated equally by the law and that all are subject to the same laws of justice, without privilege or discrimination. When talking about the digitization of legal systems the concept of legal equality often comes up. This is due to the fact that we always have to make sure that any changes made to an existing legal system do not interfere with the principle of legal equality.

In most cases, more digitization and automation mean more legal equality since a computer, AI or an algorithm is fully unbiased, unless it is programmed otherwise.

When it comes to binary translation of case law and legislation what really matters in respect to legal equality is what is being done with the outcome of this translation. As discussed before, the main reason for binary translation of legislation and case law is its potential use, for example in systems like TAX-I. It is important to look at the effects binary translation could have in the future if it would be fully implemented. One of these potential effects is increased legal equality.

If case law and legislation would be written and published in a way that a computer could understand, it would allow for a more digital and automated legal field. Examples of this legal field of the future could be increased use of automated decision making systems, artificial intelligence that tries to predict the outcome or chance of success of a case and even AI lawyers. At first glance it might seem great to have a legal system with minimal chances of human error, high efficiency and low labour costs, however there are also risks connected to this. The first danger that comes to mind is the fact that behind these innovative systems are algorithms. These underlying algorithms embody a considerable potential for discrimination and unfair treatment. The main thing that matters to these algorithms is the group the individual is assigned to. Discrimination is one of issues the principle of legal equality tries to avoid. Therefore, a potential risk of binary translation of case law and legislation is discrimination.

If the use of artificial intelligence and automated decision making systems is increased as a result of binary translation of case law and legislation, the inherent risk of discrimination also increases.

Thus, to conclude, the effect binary translation of case law and legislation will have on legal equality depends on what it will eventually be used for. If AI will use this newfound data to form patterns and try to predict the future based on case law there is a serious risk of profiling and discrimination. However, if the algorithms that drive these programs are written in such a way that it does not allow for any discrimination, it could have a positive effect on legal equality since the algorithms work the same way every time, irregardless of the subject.

Limitations of the binary translation of case law and legislation. Now that we have discussed the possible benefits of the binary translation of legislation and case law we shall examine possible limitations of binary translation.

To make the binary translation of case law and legislation possible it is necessary that the knowledge of programmers and other kinds of technicians is used. This leads to another kind of legal professionals besides the traditional lawyers, attorneys. The question that arises is if these new legal professionals are bound by the same rules and laws as lawyers and attorneys, according to Van Oostrom-Streep. You could think that these new legal professionals should be bound by the same ethical and legal rules that apply to lawyers and attorneys.

Besides the technological aspect of implementing a system of automatic binary translation of case law and legislation, one of the main limitations in our opinion is the problem of monotony, in other words, a lack of diversity. If the future of the field of legal advice is based on the use of algorithms and AI there is a serious chance it will become a matter of which firm has the best AI instead of the better lawyers/advisors. At a certain point the AI and the algorithms of these law firms will be optimised and they will most likely not be very different from each other. This, in turn, will result in very similar legal advice from many different firms since they all use the same inputs for their algorithms, being the binary translations of legislation and case law. The aspect of human interpretation of legislation and case law will be lost in the process.

This technological competition of course has its advantages for citizens because it will result in faster, cheaper and potentially better legal advice as these are common results of competition in a certain market. The main question that arises here is whether this is desirable. It might be beneficial for the consumer in the short run since access to proper legal advice will be made easier and less expensive. However, there is a chance that legal innovation might stagnate when legal advice is mostly done by AI and algorithms since human interpretation will only play a minor role in the process.

Another issue with the binary translation of legislation and case law is the fact that it will be very difficult to include every aspect of the decision into the binary code. There is a chance some of the aspects that led a judge or lawmaker to a certain decision will be lost in the translation from human language to binary language. This is due to the fact that there are many variables in play in such situations. For example the economic and financial situation of the subject, the severity of the situation and many other personal circumstances.

There is a surprising amount of similarities between the field of Law and AI as they share method and subject matter. In both law and AI there is a need to find the right balance between the order of the formal and the chaos of the informal. In law, rules have exceptions, reasons are weighed and principles are guiding. In AI, reasoning is uncertain, knowledge is context-dependent and behavior is adaptive.

According to Verheij the common subject matter in the two fields is the coordination of human behavior. There is a difference in how the respective fields tackle this problem, being that in AI “such coordination is steered by the elusive tool of intelligence” whereas in law the rule of law is the primary tool for coordination. This could pose a problem for the future of the cooperation of law and AI since AI focuses mainly on the roles of knowledge, reasoning and interaction in coordination while the legal field also takes into account how contracts, punishment compensation and authorities guide human society. For the near (and distant) future, the question remains if AI will be able to comprehend the effects case law and legislation can have on society and include these effects in its decision making like humans can