The American Journal of Political Science Law and Criminology
36
https://www.theamericanjournals.com/index.php/tajpslc
TYPE
Original Research
PAGE NO.
36-39
10.37547/tajpslc/Volume07Issue06-07
OPEN ACCESS
SUBMITED
12 April 2025
ACCEPTED
08 May 2025
PUBLISHED
17 June 2025
VOLUME
Vol.07 Issue06 2025
CITATION
Hurshid Sadikov. (2025). Predictive Justice and Artificial Intelligence:
Comparative Legal Analysis of Judicial Practice. The American Journal of
Political Science Law and Criminology, 7(06), 36
–
39.
https://doi.org/10.37547/tajpslc/Volume07Issue06-07
COPYRIGHT
© 2025 Original content from this work may be used under the terms
of the creative commons attributes 4.0 License.
Predictive Justice and
Artificial Intelligence:
Comparative Legal Analysis
of Judicial Practice
Hurshid Sadikov
Independent applicant of the Supreme School of Judges, Uzbekistan
Abstract:
The article is dedicated to research on the
implementation
of
artificial
intelligence
(AI)
technologies in the justice sector, with a focus on
predictive coding and "predicted justice." A
comparative legal analysis of the practice of the USA,
China, European Union countries, and France was
conducted. Special attention is paid to the role of AI at
the proof stage: analysis of evidence, automation of
legal procedures, and reduction of corruption risks.
Examples of court cases illustrate both the possibilities
and risks of applying AI in court proceedings.
Institutional and ethical limitations, including issues of
trust, algorithm reliability, and equality of parties, are
also considered. It is noted that despite technological
progress, full confidence in AI in the judicial system has
not yet been formed. It was concluded that it is
necessary to develop general regulations and gradually
implement AI, taking into account the principles of
procedural fairness.
Keywords:
Artificial intelligence, predictive justice,
evidence, court, predictive coding, judicial practice,
algorithms, trust, ethics, digitalization, blockchain,
information disclosure, justice, technology, comparison.
Introduction:
The development of artificial intelligence
(AI) technologies is increasingly impacting various
spheres of public life, including justice. In today's digital
society, there is a need to increase the efficiency and
objectivity of judicial processes, making it relevant to
study the role of AI in these aspects. The application of
artificial intelligence for the analysis of evidence and
forecasting of court decisions is of particular interest,
where the accuracy and neutrality of machine
algorithms can both contribute to fairness and create
certain risks.
The American Journal of Political Science Law and Criminology
37
https://www.theamericanjournals.com/index.php/tajpslc
The American Journal of Political Science Law and Criminology
This issue is increasingly attracting the attention of
researchers and practitioners. Today, the British
philosopher Andy Clarke's "predictive coding concept"
is considered one of the most important scientific
discoveries[1]. The predictive coding method of
computer technologies and information processing
algorithms allows for the analysis of large unstructured
data volumes while significantly reducing time costs.
Arguments in favor of predictive, "predicted" justice
involving artificial intelligence (hereinafter - AI) are
increasingly being voiced.
There are not many court decisions in the world that
have been made based on the conclusions of the
program using artificial intelligence. Da Silva Moore's
case against Publicis Groupe appears to have been the
first case of using artificial intelligence to evaluate
evidence[2]. It was resolved in the USA (New York) in
February 2012. The employer was accused of building
a "glass ceiling" for the women who worked for him.
To issue a decision, it was necessary to examine more
than 3 million electronic documents stored by the
defendant. The respondent suggested using the
predictive coding method. The judge reviewing the
case, Andrew Peck, accepted this proposal.
Interestingly, the plaintiffs, whose interests were
supposed to be served by the computer program,
expressed their disagreement with this and appealed
the decision to the district court. Their arguments were
as follows: the judge relied excessively on external
documents; the defendant's expert is biased, as the
chosen method of evidence assessment will benefit
him; the judge did not conduct evidence hearings
properly; the judge used the version of the protocol on
computer disclosure proposed by the defendant.
These arguments are set forth in the district court's
decision[3].
There are cases where the court initiates and insists on
the use of artificial intelligence technologies contrary
to the positions of the parties. This was the case in
EORHB, Inc. v. HOA Holdings, LLC (USA) [4]. Initially, the
judge obligated all participants to use a computer
program in the process of electronic document
disclosure and, first of all, to agree on a single software
provider. Admittedly, this requirement was amended
at the request of the plaintiffs. Ultimately, the court
amended the original ruling, agreeing with the
arguments of both the defendants and the plaintiffs.
The first allowed them to contact the software
provider and engage predictive coding, the second - to
use traditional methods of disclosing their documents.
The ratio between the small number of documents the
plaintiffs were required to disclose and the cost of the
software, which "would outweigh any benefit" from
artificial intelligence, was taken into account. This case
is interesting because the court (1) changed its original
position, (2) allowed the simultaneous disclosure of
evidence using different methods, (3) took an active
position, insisting on using predictive coding as the main
method, and (4) softened its position only after
examining the well-founded arguments of one of the
parties. The decisive argument in favor of maintaining
traditional methods for the plaintiffs was the ratio of the
final result and its cost. Note that there were no
questions about the degree of reliability of one of the
methods. This once again confirms that artificial
intelligence at the proof stage is often considered not so
much from the standpoint of its reliability as from the
standpoint of its financial cost. Obviously, predictive
coding in this case was considered a convenient tool,
allowing the parties to carry out time-consuming and
costly actions while reducing other possible costs.
Today, in China, relevant digital tools are used to
disclose and evaluate evidence as one of the procedural
stages. We are talking about an intelligent system for
analyzing evidence within the framework of online court
proceedings (blockchain plus artificial intelligence,
cloud data, etc.). Upon presentation of evidence by the
parties, this intellectual system conducts their analysis
and comparison, while simultaneously forming a list of
necessary evidence used by judicial practice in general
for similar cases. Accordingly, additional evidence not
submitted by the party (incorrectly uploaded or not
meeting the requirements) may be automatically
requested. This has especially facilitated the activities of
judges in the consideration of disputes in the financial
sphere, when it is necessary to make many complex
calculations, to give the judge the basis for considering
the case and making a final decision (Sheremetyeva,
Baturo & Y SH, 2020: 160). This opportunity arose partly
due to the use of distributed ledger technologies. The
documents uploaded to it are anonymized, marked, and
stored in cloud storage. Their analysis is carried out
using AI technologies and big data. If at the beginning of
the internet courts' functioning, only those cases that
were considered by the internet court were uploaded to
the storage facilities, later other court decisions joined
them, effectively eliminating the problem of variation in
decisions on identical cases.
The introduction of blockchain technology into the
judicial system was carried out in stages, first in test
mode in individual courts, then a unified technology for
all courts was created. In all cases, the state cooperated
with Chinese technology giants, primarily Alibaba Group
Holding Ltd (including through subsidiaries). Thus, in
2018, the "Court Blockchain" program was launched in
Hangzhou. In 2019, similar services were launched in
Beijing (March) and Guangzhou (April). In the same year,
the national "Unified Judicial Blockchain Platform of
The American Journal of Political Science Law and Criminology
38
https://www.theamericanjournals.com/index.php/tajpslc
The American Journal of Political Science Law and Criminology
People's Courts" was announced to connect all courts
of the country. It is announced that the
implementation of the platform has allowed parties,
legal entities, to significantly save by confirming the
authenticity of electronic evidence, the placement of
which in the system costs them 1 yuan (unlike the
traditional notarial certification with an approximate
cost of 4 thousand yuan). Admittedly, there are also
comments regarding information storage technology
and the trust of the courts. The technological problem
is related to the involvement of private companies in
the creation of the system, which can potentially act in
their own interests; and the thinking of traditional
judges is changing very slowly - from June 2018 to
December 2019, they recognized blockchain-
preserved electronic evidence as acceptable only in
400 cases (Wang, 2021).
In the online courts of the PRC, the consideration of
cases is limited to the subject area - we are talking
about offenses on the Internet when carrying out
online trade, a number of financial transactions, when
resolving disputes on copyright, essentially all disputes
related to interaction on the Internet. The jurisdiction
of these courts is exceptional - the parties cannot
refuse to hear the case in this court if their case falls
under the jurisdiction of an internet court. Unlike, for
example, the internet courts in South Korea, where the
consent of both disputing parties is mandatory. Court
proceedings are conducted entirely in digital format,
starting with the submission of materials to the court,
including conducting court sessions and issuing
decisions with the participation of artificial
intelligence.
It's worth noting that despite discussions about China's
transition to a "smart courts" system in the context of
using a new court model within the intellectual judicial
system, researchers still note that there is no such
digital judge issuing AI decisions. "Smart Court" is
aimed not so much and not at all at replacing a judge
with AI, but at minimizing corruption risks and ensuring
sound court decisions. This was done not to replace a
live judge with an electronic one, but rather to reduce
corruption and unfounded decisions [5]. Thus, the
decisions of the Beijing Internet Court are ensured by
twenty-nine judges, chaired by Jiang Ying, who, in
addition to diplomas of legal education (bachelor's and
master's degrees), also have an engineering education
(bachelor's degree) [6].
Although cautiously, predictive practices are still
prevalent in European litigation. There are interesting
projects of the European Union related to "predictive
justice" (predicting justice), where algorithms are used
to analyze a multitude of cases in a short time using
artificial intelligence (AI), which allows, to a certain
extent, to anticipate the outcome of the dispute
(Biryukov, 2019). The European Commission on the
Effectiveness of Justice (CEPEJ) of the Council of Europe
approved the "European Ethics Charter on the Use of
Artificial Intelligence in Judicial Systems and Their
Environments" (December 4, 2018) [7]. From the
content of the Charter, it follows that judges in the
member states of the Council of Europe do not often use
predictive tools for forecasting, although a number of
studies have been conducted.
Thus, at the initiative of the French Ministry of Justice,
two appellate courts in Rennes and Douay in the spring
of 2017 agreed to test the predictive justice software in
various court appeals, using it as an experiment in the
consideration of civil disputes, since the criminal case
was excluded from the scope of the experiment for
ethical reasons: civil, social, and commercial decisions of
all French appellate courts were analyzed. A three-
month trial was conducted using software designated
by the panel of judges as "predictive." It was proposed
to assess the value of the quantitative (innovative)
analysis of the amounts allocated by the two courts, in
addition to the geographical classification of
discrepancies noted in similar applications and tests.
The purpose of the software was to create a decision-
making tool capable of reducing, if necessary, their
excessive variability in the name of the principle of
equality of citizens before the law. The results of the
experiment were controversially discussed by two
appellate courts, the Ministry of Justice, and LegalTech,
the company that developed the product. On October 9,
2017, the Ministry of Justice and the First Presidium of
the Rennes Court of Appeal, emphasizing the modern
approach, found the software "not particularly valuable
for judges," as "high-quality tools for analyzing judicial
practice in cassation and appellate courts" already exist.
Moreover, it was indicated that the statistical approach
dominates in the software due to the detriment of
qualitative analysis and, in some cases, the fixation of
erroneous results. Indeed, unlike the Anglo-Saxon
system, the French legal system is not built on the
system of precedent law, and court decisions are made
based on "precise analysis of facts for each case"
without connection to previous decisions (Rozec &
Thiebaut, 2017).
It is clear that EU member states are attempting to
implement the idea of "predicted justice" at the national
level using predictive technologies/tools. In this context,
the European Parliament and the Council of Europe
established a single digital gateway for cross-border
evidence exchange and the procedure for handing over
judicial and extrajudicial documents (requests,
confirmations, receipts, certificates, and notifications)
in civil or commercial cases by the Regulations
The American Journal of Political Science Law and Criminology
39
https://www.theamericanjournals.com/index.php/tajpslc
The American Journal of Political Science Law and Criminology
2018/1724 (October 2, 2018) and 2020/1784
(November 25, 2020). It is believed that this should
increase the speed of transfer of both judicial and
extrajudicial documents in transnational civil
proceedings [8].
Thus, when deciding on the possibility of using
predictive coding as a tool for predicted justice, judges
adhere to certain rules. They try to obtain the consent
of the parties to disclose evidence using such
programs, even if they themselves initiate this process.
In the protocols regulating the disclosure procedure,
the parties are instructed to assist each other.
Verifiability of data plays a significant role. Since the
participants in the process, like the court, are difficult
to verify the reliability after the program issues a
verdict, much attention is paid to the stage preceding
the start of the program's work, including the
development of general rules for disclosing
information and the specifics of machine learning for a
specific case. This ensures confidence in the future
result. If the parties cooperate with each other, if each
of them has access to information, if the court
reasonably responds to the demands, objections of the
participants, then it will be difficult for them to dispute
the results of the disclosure of evidence derived by
artificial intelligence, since everyone had equal
opportunities to participate in the process and
influence the result.
We note that in a process where public and private
interests compete, these computer tools are not so
widely applicable anywhere, due to the inequality of
the parties, when it is more difficult to agree on a single
method, having previously overcome mutual distrust.
Although in civil proceedings, artificial intelligence has
not yet become the dominant method. The golden rule
of proof remains the manual human verification of
documents. This can be explained by the fact that this
software has not yet gained full trust among the
general legal community. The reliability of decisions
made during the proof process is also people's
willingness to trust the court. Thus, the European Court
of Human Rights indicated the inadmissibility of
seeking protection from a court to which the applicant
has completely lost trust. To earn comparable levels of
trust in artificial intelligence, it takes time and a
successful history of its use to solve legally significant
tasks in various spheres of human life.
CONCLUSION
The
conducted
comparative
legal
analysis
demonstrates that the use of artificial intelligence (AI)
technologies in judicial systems has become a relevant
aspect of digital transformation in justice. The
practices of the United States, China, European Union
countries, and France illustrate diverse models of AI
integration
—
from predictive coding and online courts
to intelligent platforms for data analysis. These
approaches help accelerate judicial procedures, reduce
costs, and potentially enhance the objectivity of
decisions. However, they also raise institutional, ethical,
and legal challenges, such as lack of trust in algorithms,
absence of universal standards, inequality between
parties, and the risk of overreliance on technology.
For Uzbekistan, the studied experience offers significant
practical value. A step-by-step implementation of AI at
early stages of legal proceedings
—
particularly in
evidence analysis and document automation
—
appears
most promising. It is advisable to develop a regulatory
framework governing the use of AI in the judiciary and
to launch pilot projects following international models.
Successful integration of AI must be supported by the
modernization of judicial digital infrastructure and the
training of specialists with interdisciplinary expertise
(law + IT). Given the national legal context,
implementation should be cautious, with a strong
emphasis on procedural fairness, transparency, and
public trust in the justice system.
REFERENCES
Andy Clark, ‘Beyond the “Bayesian Blur”: Predictive
Processing and the Nature of Subjective Experience’,
Journal of Consciousness Studies, 25 (2018), 71-87.
Moore v. Publicis Groupe, 287 F.R.D. 182 (2012)
Available
at:
https://www.casemine.com/judgement/us/5914efbaa
dd7b0493496dd52
Da Silva Moore, et al. v. Publicis Groupe, No. 11-Civ.-
1279 (ALC) (AJP), 2012 WL 607412 (S.D.N.Y. Feb. 24,
2012)
Available
at:
https://law.justia.com/cases/federal/district-
courts/new-york/nysdce/1:2011cv01279/375665/96/
EORHB, Inc. v. HOA Holdings, LLC, No. 7409-VCL, 2013
WL 1960621 (Del. Ch. May 6, 2013) Available at:
https://www.wilsonelser.com/writable/files/Legal_Ana
lysis/389292-v1-eorhb-v-hoa-holdings-no-7409-
vcl-
2013-esi-update.pdf
В Китае внедрили судебный ИИ. Или нет? Available
at: https://habr.com/ru/post/677920/
Judges:
Beijing
Internet
Court.
Available
at:
https://english.bjinternetcourt.gov.cn/judges_3.html
European Commission for the Efficiency of Justice
(CEPEJ) European ethical Charter on the use of Artificial
Intelligence in judicial systems and theirenvironment.
Available at: https://surl.li/ombpwg
Цифровизация судебного процесса: опыт Евросоюза.
Available
at:
https://habr.com/ru/company/
digitalrightscenter/blog/696846/
