By María Jesús González-Espejo and Ebru Metin
On the 9th February, a new Innovation in Law Studies Alliance webinar took place (access here to the recorded session: https://youtu.be/QETTmnoPua4). On this occasion the speakers have shared their experiences about expert systems (also referred as Rule Based Systems or Decision Support Systems) and what they can contribute to the legal sector. We would like to thank Ivar Timmer, Professor of Legal management & technology at the Center of Expertise Applied AI/ Legal Tech Lab/ Amsterdam Research Center for Societal Innovation at the University of Applied Sciences in Amsterdam, as well as Tomer Libal, Professor of Computer Science at the American University of Paris and principal investigator of the icomplai project (https://www.legai.app/) of the University of Luxembourg, for their generosity in sharing with us their visions on expert systems applied to the legal field.
In this article, we summarize what we had emphasized at the webinar and explain what this category of artificial intelligence named as expert systems consists of and why organizations and professionals in the legal sector should pay attention to it. We especially refer to the large municipalities of Spain and business consultancies and law firms with repetitive and high-volume work.
What are expert systems?
When we talk about expert systems, we are referring to a category of artificial intelligence that has been developing since the 1960s, which is mature and consolidated and has gone through different phases of popularity. This different perception of expert systems has been largely due to the difficulty that their development traditionally entailed and that is above all a consequence of the fact that they are based on the knowledge of people, who will necessarily have to take part in their development.
In addition to this, it is necessary to structure that knowledge in such a way that the machine can process it and produce results which requires the development of efficient methods and technology to do so. Something that has already been achieved today, for example, by the team led by Ivar Timmer from Amsterdam.
What is the difference between expert systems and machine learning?
The main difference between these two categories of technology is that while the former is based on the knowledge and experience of experts, the latter is based on data. As a consequence, the challenges generated by each type of solution are different. While in the former the difficulty derives, as we have previously pointed out, from the need to have the best experts, in the latter, the challenge is to obtain the best data. If the data is not of quality, the solution can incur problems such as the famous bias.
On the other hand, we are all already aware that artificial intelligence falls within the categories of high-impact technology, due to its transformative power and capabilities. For this reason, it is necessary that AI solutions respect a series of ethical and legal principles such as, among others, explainability, transparency or privacy. In the case of machine learning, there is a greater risk of not respecting these principles, while introducing these principles during the design phase in an expert system is relatively simple.
Why are expert systems so interesting for the legal sector?
Well, very simple, because most legal professionals have more work than they can afford, they lack time. Much of this work is routine, with little added value.
Another reason that makes them interesting is the phenomenon of hyper-regulation. It is impossible to assimilate so many rules. Neither the citizen nor the professional is capable of keeping up to date, understanding and applying a legal system that has become an unreachable jungle. Expert systems can do it, if we teach them to understand the rules, they can guide us, teach us to apply them and even make the most lawful decisions and they will do all this, in an explainable way, because we will have been the ones who told them how to do it. In addition, expert systems can help us improve the quality of standards by requiring representation according to computer-executable logic, which helps clarify and simplify natural language, as well as identify and remove unwanted ambiguities and inaccuracies.
Another reason is that our sector is essentially based on legal knowledge which is in the head by a few people. Knowledge of the regulations, of the jurisprudence, of the opinion of the public administrations and of the doctrine. The legal sector is based much more than on data, on a wealth of knowledge.
On the other hand, expert systems offer some advantages that are very beneficial, such as consistency (since the knowledge has been transferred to the solution that will always use it in the same way), legal certainty (the user receives an answer from which there is constancy) or the immediacy of the response (machines do not have schedules or vacations and are capable of processing solutions much faster than human beings).
In addition to these, expert systems are interesting because they enable legal professionals to modelize the legal knowledge autonomously, without the need to rely on a computer scientist in full. In a way a legal professional can build the knowledge base by him/herself [1].
For which specific areas are expert systems most interesting?
They are ideal for those areas where important volumes of work are handled (call centers, decision centers such as legal services that resolve claims) and where the tasks are not complex, the so called, “commodity legal matters”.
There are already many real examples of the use of expert systems in the legal field. An interesting one is described in this article, published in 1986 by a team of 6 academics from Imperial College London, in which they described how they had developed an expert system, capable of deciding which people could receive British nationality. [2]
For its part, the municipality of Utrecht is using them to draw up contracts related to real estate assets and that of Amsterdam to help citizens assess whether or not it pays to appeal certain administrative decisions, as well as to guide their officials in the complex regulatory world related to public procurement. Thanks to these solutions, officials stop working on tedious, repetitive tasks that do not add value; disparities in criteria have been identified when applying the rules and responses and solutions have been unified.
Other examples include expert systems for formal consultancy, damage estimation in litigation, automated document generators, intelligent systems for disputes and computer-aided learning in the legal field [3].
The future role of expert systems
The future of expert systems should be bright. The progressive improvement of the methodologies to legislate and do so, taking into account the role that digitization also plays in regulations, will make it increasingly easier to introduce the necessary knowledge into software solutions to offer advice and guidance services and solutions that make decisions. Just as the health sector has dedicated many resources to the creation of protocols that guide the doctor when it comes to diagnosing and curing, the legal sector can equip itself with its own protocols, standards that will greatly facilitate the work of “computerizing” the legal system.
With expert systems, legal professionals will thus be able to dedicate their time to more complex tasks for which they are really necessary and will live with less stress, which will also allow them to provide a better quality service.
Expert systems, that still somewhat unknown category of artificial intelligence solutions, can replace lawyers, but so what? It would not make all the sense that we could concentrate on what is truly important.
Maria Jesus Gonzalez-Espejo is CEO of the Institute of Legal Innovation and Ebru Metin is founder of Legal Design Turkey
- Thomasset, Paquin, Expert systems in law and the representation of Legal Knowledge: Can we isolate it from the Why and the Who? http://www.lcpaquin.com/publi_pdf/Florence_89%20.pdf
- Sergot, Sadri, Kowalski, Kriwaczek, The British Nationality Act as a Logic Program, https://www.researchgate.net/publication/234805335_The_British_Nationality_Act_as_a_Logic_Program
- Cornelia, Murzea, Alexandrescu, Repanovici, Expert systems with applications in the legal domain, accessed 13/02/2022 https://www.sciencedirect.com/science/article/pii/S2212017315001619