Justice of the Future: How Artificial Intelligence Is Changing the Legal Process. An article by Oleksandr Kvasha.

Just a few years ago, talking about a ‘robot judge’ was more a matter of speculating about the future than discussing specific models of judicial reform. Now, however, the use of artificial intelligence in the judicial system is being discussed at the level of national strategies, pilot projects and professional debates. At the same time, it is important to emphasise that this is not about replacing the system with artificial intelligence, but about a gradual yet profound transformation of the entire judicial process under the influence of AI. It is not only the form but also the logic of judicial proceedings that is changing: how claims are prepared and filed, how evidence is gathered and assessed, how legal positions are formulated, how a judge reaches a decision, and how citizens interact with the court at various stages.

Artificial intelligence is already becoming a kind of ‘hidden infrastructure’ of the justice system — from the automated processing of documents and the search for relevant case law to the prediction of likely outcomes of disputes. This process unfolds differently in various countries: in some, AI is integrated into the courts’ internal information systems as a tool to assist judges; in others, online tribunals are being established, where most proceedings take place without the traditional courtroom setting. However, what they have in common is that the judicial system can no longer treat AI as a mere optional extra. It is a factor that is already influencing working standards, public expectations and even the very understanding of what constitutes a fair trial.

Therefore, the key question today is not so much whether it is permissible to use artificial intelligence in the justice system, but rather under what conditions, and with what safeguards and restrictions, this should take place. It is important to define the framework within which AI can be integrated into the judicial process in such a way as to enhance the efficiency and predictability of justice, without undermining public trust in the courts, the independence of judges and fundamental human rights.

In recent years, the foundations for such changes have been laid in Ukraine – the digital infrastructure of the judicial system. The Unified Judicial Information and Telecommunications System is already operational, as are the electronic office and the ‘Electronic Court’ subsystem; it is also possible to participate in hearings via video link. For a lawyer, this means that a claim, application or motion can be filed online, procedural documents are received electronically, case files can be accessed via the internet, and in many cases there is no need to physically travel to court. In other words, we already have an ‘e-court’ in the sense of ‘digital document management and remote participation’, but we do not yet have a court where AI actually influences the analysis of a case or the content of a judgement.

It is important to distinguish between two levels here. The first is the e-court itself. Its main purpose is convenience and accessibility: the ability to avoid carrying papers to the court registry, to submit documents from the office or home, and to join a hearing online. The second level is the ‘smart court’, where AI-based systems help to search for relevant case law, analyse large volumes of documents, suggest possible procedural decisions, and, in simple, routine disputes, partially automate their adjudication. Ukraine has, in effect, completed the main stage of its transition to an electronic court system. A far more complex question is whether we will dare to move from purely technical digitalisation to the genuine integration of AI into the substance of the judicial process, rather than limiting ourselves to its outer shell.

For example, China is taking a more radical approach. The country operates specialised online courts that hear disputes relating to e-commerce and digital services entirely online, and is also developing a system of ‘smart courts’ – courts where AI is used to search for case law, analyse documents and draft decisions. At the level of state policy, it has been declared that artificial intelligence is to become a standard working tool of the courts. In other words, digitalisation there is already about the substance of the courts’ work, rather than merely the possibility of ‘filing a claim online’.

Canada offers a different, but no less instructive, model – the Civil Resolution Tribunal in the province of British Columbia. This is a government body that handles minor disputes (between neighbours, consumer disputes, small debts and, occasionally, road traffic accidents) entirely online. A person visits the website, answers questions, and the system helps them formulate their claims and suggests settlement options. If no agreement can be reached, the case is referred to a human adjudicator — a member of the tribunal. In this model, artificial intelligence does not pass judgement, but helps the person navigate the path from problem to solution: it offers guidance, structures the information and suggests courses of action. The main focus is on facilitating access to justice for those for whom a traditional court is too expensive, too complicated or geographically inaccessible.

Looking at these examples as a whole, Ukraine appears to be a country that has already taken an important first step — moving the courts to a digital format — but has not yet decided on the purpose or method of using artificial intelligence in the judicial process. At the same time, several key areas are already clearly emerging globally where AI is being effectively integrated into the justice system.

The first area is assistance in preparing procedural documents for prosecutors, lawyers or judges. Modern AI systems are capable of suggesting the structure of a statement of claim, an indictment, an appeal or a cassation appeal, select and summarise relevant legislation and case law, and draft individual sections of a document — such as the statement of facts, legal grounds and the relief sought. They can identify formal shortcomings: the absence of mandatory details, incorrect references to sections of the law, and inconsistencies between different parts of the text. In some jurisdictions, the system effectively guides the user step by step: it asks questions, uses the answers to generate a ‘framework’ for the document, and then offers to let the user edit it. For a practising lawyer, this is not a substitute for professional judgement, but a tool for saving time on routine tasks and a way to reduce technical errors.

The second important area is the use of artificial intelligence for working with evidence and large datasets. Here, AI is used as a tool for searching for relevant decisions amongst thousands of previously issued court rulings across similar categories of cases, for the preliminary classification of documents (contracts, correspondence, financial reports, minutes, etc.), identifying typical ‘template’ wording and potential contradictions in testimony, and processing large volumes of information in complex corporate, antitrust or corruption disputes. In some countries, systems are already being tested which analyse electronic correspondence or financial transactions and highlight passages that may be of evidential value. This does not relieve the judge of the duty to independently assess the relevance, admissibility, reliability and sufficiency of the evidence, but it does make it possible to grasp the structure of the case file more quickly and to focus human attention on potentially key aspects.

The third area of application relates to the drafting of court decisions and the standardisation of judicial practice. In a number of legal systems, tools are already in use that propose a standard structure for a decision (introductory, descriptive, reasoning and operative parts), automatically insert standard wording for similar situations, select references to relevant legal provisions and judgements in similar cases, and compare the draft judgement with previous case law, showing the extent to which the chosen approach complies with or deviates from established standards. Functionally, it resembles an advanced word processor with a legal memory: the system suggests to the judge how arguments have typically been formulated in similar cases, but does not deprive the judge of the ability or duty to deviate from the template if the specific circumstances so require.

At the current stage of technological development, artificial intelligence is already capable of performing a significant portion of the intellectual support work for participants in criminal and civil proceedings. In particular, this involves the preparation of draft procedural documents for prosecutors, attorneys or judges: indictments, statements of claim, appeals and cassation appeals, applications, and drafts of specific sections of court decisions. AI systems can rapidly process large volumes of legislative material and case law, carry out an initial legal assessment, suggest a document structure, wording for legal arguments, and also draw the user’s attention to gaps in the reasoning or the absence of necessary details. In fact, artificial intelligence can already serve as a tool for the automated generation of legal drafts, which are subsequently refined by a lawyer taking into account the factual circumstances of the case and the chosen legal strategy.

Read the full article on the ‘Yurydychna Praktyka’ website via the link.