276

The purpose of the research is to analyze the use of artificial intelligence in the justice systems

of developed countries and the prospects for its use in criminal proceedings in Ukraine.

Methodology: Research from the analysis of documentary sources; the basis is the dialectical

method of knowledge of the facts of social reality, on which the formal legal and comparative

legal approaches are largely based. Conclusions. In the near future, the accompanying

efficiency and simplifies the procedural decision-making procedure in terms of economy.

KEY WORDS: artificial intelligence; automation; computers; cybernetics; Legislation.

*Professor, Doctor of Science in law, Professor at the Department of Administrative and Customs Law,

Recibido: 01/10/2021 Aceptado: 29/11/2021

277

El objetivo de la investigación es analizar el uso de la inteligencia artificial en los sistemas

judiciales de los países desarrollados y las perspectivas de su uso en procesos penales en

Ucrania. Metodología: Investigación a partir del análisis de fuentes documentales; la base es

el método dialéctico de conocimiento de los hechos de la realidad social, en el que se basan

en gran medida los enfoques jurídicos formales y jurídicos comparados. Conclusiones. En un

futuro próximo, las medidas organizativas de acompañamiento para la implantación de la

crea un modelo de automatización de decisiones digitales. Dicha automatización simplifica

economía.

PALABRAS CLAVE: inteligencia artificial; automatización; informática; cibernética;

be recognized in the nearest future, accompanying organizational measures for introduction

of artificial intelligence and its regulatory support in public authorities, which are associated

with the storage of big data, information processing based on mathematical algorithms and

decision-making based on artificial intelligence will be an integral component of life of our

mathematical apparatus.

278

includes a large number of variables unknown in advance. Parameters are configured

gradually, during the learning phase, which uses learning data sets for search and

classification. Various methods of machine learning are chosen by developers choose

depending on the nature of tasks to be performed (grouped). These methods are usually

classified into three categories: supervised learning (learning with a teacher), uncontrolled

learning (learning without a teacher) and reinforced learning (learning with reinforcement).

These three categories group different methods, including neural networks, deep learning,

etc.

We can also talk about work of artificial intelligence in the sphere of activities performed by

prosecutor and police. Therefore, practice of using mathematical algorithms and software

that processes big data in the sphere of criminal proceedings with further making proceeding

decisions, requires a substantial research.

criminal law and legal regulation of artificial intelligence were researched by Sartor, L. Karl

Branting “Judicial Application of Artificial Intelligence” (Sartor, 1998), G. Hallevy “When

Robots Kill: Artificial Intelligence under Criminal Law” (Hallevy, 2015), K. Brennan-

Marquez, S. Henderson “Artificial Intelligence and Role-Reversible Judgment”, Ronald J.

279

280

criminal cases, risk assessment);

LexisNexis (analytics), COMPAS – corrective offer Management Profiling for alternative

sanctions are used by US courts to assess likelihood of defendant’s recidivism and to analyze

previous misconducts;

used in the Supreme People’s Court of the People’s Republic of China.

281

imprisonment for a certain period of time (Law of Ukraine, 2015). According to parts 1, 2 of

Art. 314–1 of the Criminal Procedure Code of Ukraine in order to provide the court with

information that characterizes the accused person, as well as for making a court decision on

the level of punishment, a representative of the authorized probation body shall prepare a

pre-trial report by court order. A pre-trial report shall be compiled concerning a person

punishment never exceeding five years of imprisonment. A pretrial report on a juvenile

accused aged 14 to 18 shall be compiled regardless of the gravity of the crime committed,

except in cases provided for by the Criminal Procedural Code of Ukraine. (Law of Ukraine,

person from the database, analyzes these data using mathematical algorithms, and makes a

report providing assessment the relevant risks. The more data, the higher accuracy of the

report is. However, if the artificial intelligence takes into account data that will be artificially

created or falsified, will be based on incorrect translation, incorrect expert’s conclusion and

282

environment dated December 3–4, 2018, which stated that ProPublica found discrimination

in the algorithm used in the COMPAS software intended to assess the risk of offense

repetition, when the judge must determine the sentence individually (Leheza et al., 2018).

the risk of rime repetition as well as for assessing work of artificial intelligence in courts. But

283

advantage of working with big data is that the whole “history” of a suspect is recorded in a

proper report (Hassani et al., 2020).

Discriminating Against the Poor,” M. Berges notes that in order to assess low, medium and

high risks the HART system uses data from 34 various categories covering age, gender and

history of offenses. These data categories contain information about the postal code. Police

officers are now removing “data” with postal codes that include the first four digits of

Durham’s indexes from the HART system (UK Police are Using AI to Inform Custodial

Decisions – but It Could be Discriminating against the Poor, 2018).

demonstrates operation of artificial intelligence in prosecutor’s office and court concerning

efficiency and procedural economy. For example, the Buenos Aires District Prosecutor’s

Office in Argentina summed up interim results of testing the Prometea software application,

which can make and issue a court decision from a number of categories of civil and