Nele Parrest
Justice of the Supreme Court


When talking about information technology and the courts, the first image that usually comes to mind is a robot holding a judge’s gavel. In other words, the discussion is over when and to what extent will a robot (artificial intelligence) replace a human judge in the administration of justice. Of course, this is not the only topic that connects information technology and the courts. It is only a fragment of one. Information technology is an everyday tool used by courts to organize their work and proceedings.

Cybercrime, contracts signed with digital tools, using algorithms in administrative tasks, the legality of automated processing of personal data – these and other such issues end up on the desks of judge’s time and again and need resolution. Therefore, information technology is an everyday part of practically every judge’s work.

Considering the above mentioned, it is understandable that the theme of the “Yearbook of Estonian Courts 2019” is the relationship between information technology and the courts. This turned out to be particularly prophetic when the COVID-19 pandemic seized the whole world, including Estonia, at the beginning of 2020. The emergency situation declared in Estonia[1] forced all of society, including the court system, to find means of communication and forms of working that avoid physical contact between people and, if possible, having to leave your home. Information technology became the primary working tool even for those who weren’t acquainted with or lacked the desire to familiarize themselves with it previously.

In 2018, the Supreme Court en banc made a remarkable judgment concerning digital issues, stating that “today, persons representing private law legal persons should generally be able to communicate with the state electronically.”[2] In other words, just like in Estonia we can expect every person to know how to read and write,[3] according to the Supreme Court, the legislature can now create legislation with the assumption that at least the representative of a legal person has digital literacy, i.e. the knowledge, skills and willingness to use a computer. Moreover, the communication in question wasn’t between a business entity and the state, where higher standards are to be expected, but in the non-profit sector – meaning this could practically affect almost every person in Estonia. The referenced quote is undoubtedly a brave and noteworthy stance on the part of the Supreme Court, which many might question as ahead of its time.

It is probably not a question of whether judges have the basic knowledge and skills to use a computer considering the generational change currently taking place. But circumstances may arise in some matters where a legal education is not sufficient to resolve a particular dispute. How can we verify the legality of an administrative act issued by artificial intelligence? Which nation’s law applies to an intermediary service that aims to profit from the use of an electronic platform to bring together potential customers and service providers when all three parties are located in different countries? Does e-voting comply with the general principles of elections? And other such questions. Thus, one is justified in suggesting that information technology be a reasonable part of a university’s faculty of law program. The law is and will always remain merely a means of resolving debates that cover various areas of society in all their diversity. A judge who deals with juvenile matters, bankruptcy law, environmental law, institutionalization of the mentally ill, etc. must inevitably educate him or herself in that respective field, whether that means obtaining psychological, medical, economic or other knowledge. The same is evidently true for information technology, where every judge must keep abreast of developments to a greater or lesser degree, depending on their specialization.

Information technology-related matters are undoubtedly an exciting challenge for judges also in the sense that legislation may tend to be so-to-say outdated. Meaning: the standards set for an analogue society must now be adapted to the conditions of a digital society. That definitely requires a creative ability to interpret and apply the standards using traditional methods of interpretation but without contradicting the general principles of law. For example, in a matter related to Estonia, the European Court of Human Rights recently stated for the first time that “access to the internet is increasingly being treated as an independent right…”[4] It is probably only a matter of time until it becomes an independent human right. Analogous examples can be found in the case law of our Supreme Court. For example, do money laundering rules apply to the virtual currency bitcoin,[5] even though bitcoins didn’t even exist at the time the law was passed? Are e-cigarettes comparable to medication?[6] Etc.

An important question for judges will be how they feel about research, analysis and even services that try to predict the views of judges in resolving matters. They will probably be rather cautious and reticent. The European Court of Human Rights has repeatedly found that extensive media coverage of one matter or another has violated the rules of justice because it may have affected the outcome of the matter.[7] Trying to predict a judge’s decision may have a similar effect. For example, knowing that the prediction says a judge should make decision A might make the judge subconsciously choose B instead to show that the prediction was wrong.

Coming back to the beginning, the answer to the question of will artificial intelligence one day replace human judges is probably not. We cannot rule out the possibility that the society of the future will have artificial intelligence at its disposal that can successfully resolve at least certain court matters and possibly even better than a human judge. But do the members of society want their problems resolved by an algorithm instead of a person? At the moment, we can safely say no. It’s not out of the question that we may one day, in certain areas, talk not about the right of access to a court presided over by a person but about the right to appeal a decision made by an artificially intelligent judge to a human judge. But no more than that.

True, just as the human mind is a “black box” that we cannot completely fathom or explain, so too is artificial intelligence. It can be argued that this means there’s no difference in the end which one makes the final judgment. But there remains one big difference. And that’s trust. Even though we might not fully understand how a human brain’s decision making process works, including in resolving a matter, a democratically legitimate parliament has given the judges of the highest court the right and obligation to say what is right, fair and ultimately true and have trusted them in this. As long as members of society lack that same trust in artificial intelligence and don’t want to permanently replace human judges, there is no reason why the legislator should provide for such a thing. Of course, in order for society to trust us, the human judges, we the entire judiciary must keep working hard every day to earn that trust.

Happy reading!


[1] Government 12.03.2020 Order no. 76 „Declaration of Emergency Situation in the Administrative Territory of the Republic of Estonia“ (RT III, 13.03.2020, 1)

[2] Supreme Court en banc Decision no. 2.10.2018, 2-17-10423/10, p. 59.1. Cf. Supreme Court en banc Decision 10.12.2013, No. 3-4-1-20-13, where the Supreme Court declared the State Fees Act unconstitutional and repealed the section that set a higher state fee for applications submitted through ohter channels than for those submitted electronically through the website

[3] According to the UN Human Development Report 2019, the literacy rate in Estonia is 99,9%, (available at: (1.03.2020)).

[4] European Court of Human Rights Decision 19.01.2016, no. 17429/10, Kalda vs. Eesti, p. 52.

[5] RKHKo 11.04.2016, no. 3-3-1-75-15.

[6] RKHKo 28.A04.2015, no. 3-3-1-5-15.

[7] Nt EIKo 24.01.2017, no. 57435/09, Paulikas vs. Leedu, pp. 57 ff.