6 minute read
"Meet your Robot Judge"
Artificial intelligence and Legal Judgement Prediction: a blessing or a threat to the effectiveness of the judicial systems? MEET YOUR ROBOT JUDGE
Angeliki Konstantara
Member of ELSA Athens
Fair administration of justice constitutes a cornerstone to state practice and policies, ensuring the protection of fundamental human rights and freedoms against any public or private infringement. Such administration requires the functioning of effective and reliable legal and judicial systems to fulfil the purpose of social stability and public order. Therefore, there have been successful attempts to integrate artificial intelligence into the judicial process concerning not only civil law cases, such as divorces, but also criminal convictions. AI programs, namely Legal Judgment Prediction (LJP)1 and Self-Attentive Capsule Network (dubbed as SAttCaps)2, are currently very popular in the US and in Europe, leading to the adoption of a legal regulatory framework for the use of AI in judicial decision making. The adoption of such safeguards is of vital importance so as to avoid the risks of technological growth to the detriment of human rights and fundamental freedoms while enhancing the effectiveness of the legal systems. Thanks to the access to legal judgement data, all the three stages of a trial (pre-trial claims, in court debate and judge sentence stage) are simulated online and legal based factors are evaluated with case life-cycle and multi-task learning mechanisms.3
1 Luyao Ma, Yating Zhang, Tianyi Wang, Xiaozhong Liu, Wei Ye1, Changlong Sun, Shikun Zhang, Legal Judgment Prediction with Multi-Stage Case Representation Learning in the Real Court Setting, arXiv:2107.05192v1 2 Yuquan Le , Congqing He, Meng Chen, Youzheng Wu, Xiaodong He and Bowen Zhou, Learning to Predict Charges for Legal Judgment via Self-Attentive Capsule Network, 24th European Conference on Artificial Intelligence - ECAI 2020 Santiago de Compostela, Spain 3 Supra note 1, 2 This multi-role representation model seems promising, yet unreliable. However time-saving, cost-effective and unbiased it is considered, it does not guarantee fairness in judgments due to lack of empathy, intuition and perceptiveness.4 Supporters of AI systems seem ignorant of the complexity of the judge’s role in the systematic and teleological interpretation of the law in the light of abstract principles, such as good faith and its application to the evidence submitted. It is certainly no exaggeration to state that the replacement of the human judge by algorithms undermine the substantial role of judicial decision making in shaping morally upright individuals, hence reinforcing social order and solidarity. But what renders AI most problematic for the fair administration of justice is the possible violation of the right to a fair trial due to the automatic adjudication of criminal cases. To be more specific, in accordance with the 6th Amendment of the Constitution of the United States5 as well as with article 6 of the European
4 Sourdin Tania, Judge vs Robot: Artificial Intelligence and Judicial Decision Making, UNSW Law Journal Volume 41(4) 5 Sixth Amendment "In all criminal prosecutions, the accused shall enjoy the right to a speedy and public trial, by an impartial jury of the State and district wherein the crime shall have been committed, which district shall have been previously ascertained by law, and to be informed of the nature and cause of the accusation; to be confronted with the witnesses against him; to have compulsory process for obtaining witnesses in his favor, and to have the Assistance of Counsel for his defence.”
Convention of Human Rights6, the criminal procedure shall encompass the right to access to a court, emanating from the rule of law, as well as the equality of arms. The European Court of Human Rights has designated the principle of equality of arms as only one feature of the wider concept of a fair trial, which also includes the fundamental right that criminal proceedings should be adversarial7 in the sense that both prosecution and defence must be given the opportunity to have knowledge of and comment on the observations filed and the evidence adduced by the other party.8 Even though AI methods seem to fulfil the requirement for an impartial and speedy trial, as both the procession of the data as well as the decision making are carried out without delay, the access to a court is compromised by the abolishment of the oral procedure. The deprivation of the defendant’s right to be heard in court might also trigger the presumption of innocence (in dubio pro reo), as the algorithms used by LJP systems might distort the meaning of a particular article or even the defendant’s written statement and thus lead to the unfair conviction of an innocent man.9 Last but not least, LJP mechanisms put the right to liberty in question, given the fact they are not sufficiently developed to carry out pre-trial assessments for the determination of a defendant’s detention pending the trial. Alarmed by these concerns, the European Commission for the Efficiency of Justice (CEPEJ) adopted in 2018, under the auspices of the Council of Europe in Strasbourg, a comprehensive European Ethical Charter on the Use of Artificial Intelligence in Judicial Systems and their environment.10 This Charter proclaims the principles that prevent the infringement of the right to a fair trial by automatic AI systems. Among these five principles, the Principle of respect for fundamental rights ensures that the design and implementation of artificial intelligence tools and services are compatible with fundamental rights and the systematic
6 European Convention of Human Rights (ECHR), article 6: “In the determination of his civil rights and obligations or of any criminal charge against him, everyone is entitled to a fair and public hearing within a reasonable time by an independent and impartial tribunal established by law.” https://www.echr.coe.int/documents/convention_eng.pdf 7 Case of Brandstetter vs Austria, (Application no. 11170/84; 12876/87; 13468/87), ECHR Judgment of 28 August 1991,par.66 8 Ibid par.67 9 Franciska Zsófia Gyuranecz, Bernadett Krausz, Dorottya Papp, The AI is now in session – The impact of digitalization on courts, a THEMIS SEMI-FINAL D – Judicial Ethics and Professional Conduct,2018 10 https://rm.coe.int/ethical-charter-en-for-publication-4-december2018/16808f699c interpretation of the law. The Principle of nondiscrimination, which is also a prominent ECHR11 principle, is major to the equality of arms, as it prohibits any discrimination that might have an unjustifiable influence on the conviction of the defendant on the basis of sex, race, colour, language, religion, political or another opinion, national or social origin, association with a national, minority, property, birth or another status. When it comes to the processing of intangible data and certified sources, the Principle of quality and security guarantees the legitimate preservation of personal and sensitive data with models elaborated in a multi-disciplinary manner, in a secure technological environment, whereas the Principle of transparency, impartiality and fairness as well as Principle “under user control” preclude a prescriptive approach and ensure the role of users as informed actors, capable of designating the scope of their data procession. It remains to be seen how all these principles will be applied in the AI judicial mechanism. In conclusion, the digital era calls for the transformation of traditional methods towards flexibility and speed. Even though AI is well adapted in automobility and engineering, its application in humanitarian and social services remains dubious, especially as far as legal administration is concerned. The replacement of the oral pleadings with multi-role mechanisms does not correspond to the defendant’s need to have a hearing in front of the judges, whereas the judgment is solely based on the sterile evaluation of legal and case facts, without any interpretive ability. Considering the educative and humanitarian role of a judge in a society, the development of AI mechanisms to the detriment of the support of the existing courts must be viewed with vigilance especially in criminal procedures, where the burden of proof is beyond any reasonable doubt, rendering the judgment decisive for a person’s life. That does not entail, however, the indiscriminate rejection of AI technology in our legal mechanisms. AI technology has a lot to offer in terms of efficiency and therefore lawyers and judges could take advantage of its benefits when dealing with identical, typical civil law cases, such as contracts or consensual divorces. In any case, it should be humans who have the upper hand in justice administration, not mere algorithms.