The legal sphere, like no other, requires changes because the problems of bureaucracy and the processing of huge amounts of data are a sore spot for any lawyer, prosecutor and judge. After all, it is quite common that one side of the trial sends the other side a list of documents for review. According to the first side, it makes sense to study all of these documents in the context of the case. So, this list may include two million documents (!) and, of course, all of them cannot be studied. Moreover, imagine what kind of headache many similar claims can create in case of a natural disaster or an industrial marriage. Now imagine how cool it would be to automate their grouping and ordering right up to merging into one class action lawsuit. And these are not lengthy thoughts of the article’s author, but quite real words of one of our clients.
The use of machine learning and artificial intelligence can solve problems of this nature. Although experts from many industries fear that machines will force them out of the labor market, the lawyers can stay calm. AI is only designed to take on routine tasks and give people free time for creative or strategic decisions. Moreover, steps in this direction are already being actively taken. For example, in 2016, the entire legal world was shocked by the news that an algorithm had been created that guessed the decisions of the Strasbourg court with a probability of up to 79%. Scientists from the United States and Great Britain who were involved in the creation of the algorithm are convinced that it will eventually be in demand in the selection and sorting of cases by importance. This effect was achieved thanks to algorithms that know which circumstances most often draw the attention of a judge when considering a case.
Investors from the USA are aware of the legal tech direction prospects.
This is an information for you to think about. In the course of this article we will consider more cases of AI usage in legal tech. In the meantime, let's not get ahead of ourselves and look at how and when machine learning has entered legal tech, why this is slower than desired, and how smart services will help make legal processes more effective in the future. The court is now in session.
What are the circumstances of the case?
What were the first examples of lawyer jobs automation? Let's be honest: jurisprudence is an area where there is a lot of routine. What programs have helped lawyers make their work easier? Of course, the very first and most important program was Microsoft Word. Its appearance in 1983 totally changed the process of legal documents preparation. Numerous text formatting options were not previously available to the audience. Perhaps, the fact that Word found a response among lawyers contributed to its popularity among millions of users later.
In the late 80s and early 90s, the next step in the development of the legal tech industry took place: reference legal systems appeared. At the same time, in the early 90s, if the articles of a law changed, then couriers with floppy disks containing new versions of documents left for the customers who bought a legal reference system.
An equally important milestone in legal tech is the release of FineReader, which can translate scanned documents into text. By the way, in the modern version of FineReader, machine learning is already in full use.
Let's move on to the present. When ML first entered legal tech? In fact, not later than in other areas, despite the conservatism of this industry and it is all due to e-Discovery - the process of legal documents exchange in a digital way.
It was the need to simplify the e-Discovery process that became the catalyst for the appearance of the first ML legal tech startups in 2010-2012. This still remains one of the key areas in legal tech along with the research and management of contracts between firms. is still - along with research and management of contracts between firms - remains one of the key areas in legal tech. The best-known startups linking law with IT have been successful thanks to their efforts in e-Discovery. Examples are Relativity from Chicago and Everlaw from California. The legal tech software and algorithms for e-Discovery are universal in their own way - similar solutions are also used in corporate and financial investigations.
So, why are there still few examples of ML applications in the legal tech industry? One of the problems is that not all documents have been digitized yet, so smart systems have nothing to train on. For example, bank employees do not always digitize loan agreements. Paper media is more difficult to store, access to them is sometimes difficult: specialists have to go to archives and take paid copies, not to mention the lack of a full-fledged search for such materials. In addition, paper documents are easier to fake.
Let’s interrogate witnesses
We will now disappoint you a little bit by repeating the thesis mentioned at the very beginning of this article. So, whether machine learning can replace a person in the near future or jurisprudence is too delicate a matter to trust it to machine? The short answer is no - at the present moment, it cannot, but there are certain achievements in this area already.
What kind of? Let's take a look.
So, in general, the legal tech field has been developing for quite some time and in the following areas:
- automation of standard legal services with legal tech;
- expansion of legal tech online services for clients;
- transition of the justice system online using legal tech;
- creation of legal tech solutions based on artificial intelligence.
What can legal tech offer to us in these areas at the current stage of its development?
Online consultation and bot lawyers
Let's start with the simplest examples. You must have already heard about the 24/7 consultation with live lawyers in online chats. So, how about their artificial counterparts? After all, the area of law is very convenient for automation.
Moreover, not only the largest companies are engaged in automation but also new services where customers can independently and online declare income, prepare a simple statement of claim, register a business, draw up a contract.
Bots in instant messengers are perfect for automating one function. Already, there are examples of bots checking the name of a company, helping to choose a tax system and requesting scans of documents or photos of a client. In addition, there are bots that can issue legal reference information highlighting keywords from a question.
Investigations and document searches
Lawyers of various specializations use legal tech software to search for laws, by-laws or court decisions (if it is a case-law). Sometimes, to prepare for a hearing, you need to study hundreds of documents, and case files are thousands of pages long. Working with technologically obsolete databases or manually searching for information takes time, which can be used more efficiently. Legal tech tools such as AI and big data solutions help streamline the process.
In September 2018, vLex launched a beta version of Vincent a “smart” assistant who analyzes legal documents in English and Spanish and knows the laws of nine countries, including Spain, the USA, Canada and Mexico. Neural network trained for 100 million legal files.
The French Doctrine project also helps in the study of legal texts. The startup created a search engine for court decisions and other legal documents. This is especially true given the increase in open data.
BookLawyer, Nalytics and Counselytics also offer search, analytics and data structuring solutions.
Another curious phenomenon of the digital age. Let’s assume there is a very serious case that needs additional funding, for example, a trial with some mega-corporation or a socially significant case. The largest corporations are served by top law firms, and few are ready to take risks of conflict with them because of significant legal costs.
How to make justice triumph? You can use crowdfunding platforms to attract investment in litigation. Investment promotion services such as the American Lex Shares or British crowdfunding service Crowd Justice will help entrepreneurs enter into a dispute with the giants. The investor covers all legal expenses and upon winning the case receives a percentage of the amount recoverable.
Verification of documents and due diligence
Legal papers are strictly formalized. Even a small mistake can cause serious problems for the company. Legal tech platforms are also capable of checking documentation, correcting inaccuracies and arranging papers.
LawGeex, which created the analytic platform, has proven its effectiveness through competition. The start-up AI system opposed lawyers with expertise in contract law. People and the machine analyzed five Non-Disclosure Agreements. 20 lawyers completed the work with 85% accuracy (on average), and artificial intelligence reached 94%. At the same time, the indicators of the best lawyer were equal to the results of the machine, and the worst of the experts managed by 67%. The machine coped with the task much faster: 26 seconds versus 92 minutes on average for lawyers.
Despite this, not all contracts can be trusted with a computer. Sometimes the price of an error is too high, but AI can already take on the work of checking standard contracts.
Technological solutions also help prepare for a hearing, find a lawyer and learn about opponents. This information is useful both for lawyers and for people who first encountered the judicial system.
Lex Machina collects data on judges, lawyers, and disputes by analyzing a wealth of legal information. The platform works with corporations to help lawyers predict the outcome of various business strategies.
AI tools allow you to choose who will represent you in court. The Premonition service not only collects data from the entire US judicial system but also assesses the risks in selecting a lawyer in each case.
“Artificial lawyers” are also capable to assist the law firm in simple tasks. For example, some part of the tasks that trainees and junior employees typically perform can be transferred to the ROSS Intelligence AI assistant. It specializes in case law - studies cases, and then provides lawyers with relevant information that is useful in preparing for a hearing
Let’s summarize. Despite the slow pace of development of the industry, it is possible and necessary to implement legal tech solutions as quickly as possible because this is a sure way to become an innovator and to gain market share earlier than others. And we, in our turn, will render you all possible assistance in this by developing the technical part.