International Journal of Law And Criminology
25
https://theusajournals.com/index.php/ijlc
VOLUME
Vol.05 Issue04 2025
PAGE NO.
25-28
10.37547/ijlc/Volume05Issue04-06
Separate Issues of Applying Artificial Intelligence in
Forensic Expertise
Abdullaev Rustam Kahramanovich
Acting Associate Professor of the Department of Criminalistics and Forensic Science Tashkent State Law University,Doctor of
Philosophy in Law (PhD), Uzbekistan
Received:
15 February 2025;
Accepted:
16 March 2025;
Published:
14 April 2025
Abstract:
The article examines the use of artificial intelligence in conducting forensic examinations, as well as the
specifics of applying artificial intelligence, taking into account the procedural significance of the expert's
conclusion. Requirements for the process of developing artificial intelligence systems for their use in forensic
activities have been proposed.
Keywords:
Forensic examination, artificial intelligence, assessment of the expert's opinion, application of artificial
intelligence in forensic activities.
Introduction:
Today, computer technologies are
increasingly penetrating all spheres of human activity.
Computer tools are widespread that allow for the
acceleration and simplification of various human
operations - writing and editing texts, creating,
correcting, and analyzing images, performing various
mathematical calculations, automatically managing
production processes, etc.
Computer technologies are also actively used in law
enforcement activities. Currently, in Uzbekistan,
systems for computer identification of license plates of
vehicles that have violated traffic rules, and systems for
determining the appearance of persons who assist in
the search for offenders are widely used.
In recent years, with the development of computers
and the increase in their capabilities, the processes
within computers have begun to be compared with
human mental activity. As a result, a new term -
artificial intelligence - appeared.
Currently, the concept of "artificial intelligence" is not
yet fully formed. A legal framework for the
implementation and use of artificial intelligence
technologies has been developed in our country. For
example, the Decree of the President of the Republic of
Uzbekistan dated October 5, 2020 No. UP-6079 "On
Approving the Strategy "Digital Uzbekistan - 2030" and
Measures for its Effective Implementation" approved
the Strategy "Digital Uzbekistan - 2030". The Strategy
"Digital Uzbekistan - 2030" defines the strategic goals,
priority areas, and medium- and long-term prospective
tasks for the development of the digital economy and
e-government of the Republic of Uzbekistan, and also
serves as a basis for the further widespread
introduction of digital technologies based on the
priorities defined in the UN Sustainable Development
Goals and the e-government development rating [1].
Currently, artificial intelligence is developing and being
widely implemented in various spheres of modern
society, including law enforcement activities and the
practical activities of forensic experts. In forensic
examination, artificial intelligence is used for the
primary processing of large volumes of data contained
in databases for the study of papillary patterns, DNA
samples, objects of ballistic examination. With the help
of artificial intelligence, large volumes of textual
information on social networks are analyzed. "It is
known that modern artificial intelligence consists of
algorithms and software systems designed to perform
various operations, and performs several tasks that
human consciousness can perform based on
information entered into the database" [3].
The discussion on the main possibility of using artificial
intelligence in law enforcement activities has
concluded. The very practice of law enforcement
agencies has confirmed the effectiveness of using
International Journal of Law And Criminology
26
https://theusajournals.com/index.php/ijlc
International Journal of Law And Criminology (ISSN: 2771-2214)
artificial intelligence. At the same time, the question
remains open about the limits, forms, and methods of
using artificial intelligence, for example, whether it is
possible to fully rely on it when conducting a forensic
examination. Currently, most authors believe that this
is impossible at the current stage of development of
legislation, expertise methodology, and artificial
intelligence.
The limitation on the use of artificial intelligence in
forensic examination is explained by the specifics of
expert research and its position in the modern judicial
process.
These features are that the expert's conclusion does
not have a predetermined force, and "the expert's
conclusion is assessed by an official of the div
conducting the pre-investigation check, an investigator,
an inquiry officer, or a court, together with other
evidence collected on the materials of a criminal case
or pre-investigation check, from the point of view of its
scientific validity and compliance with all procedural
rules established for conducting the examination" [2].
In particular, the expert opinion is assessed in
accordance with:
- compliance with the established procedural
requirements for the appointment and conduct of the
examination;
- the reliability of material evidence submitted for
expert research, their sufficiency to resolve the issues
posed to the expert;
- the expediency and scientific validity of the research
methods and methodology chosen by the forensic
expert for the case under consideration;
- completeness and comprehensiveness of the research
conducted by the forensic expert, compliance of the
research with the chosen methods and methodology;
- logical justification of the intermediate and final
results of the research;
- the correspondence of the expert's final answers
(conclusions) to the evidence available in the case.
The most difficult task for evaluating an expert's
conclusion is related to the research methodology.
When it comes to the assessment of traditional forensic
examinations conducted using an approved and
scientifically based methodology, this, of course, does
not present much difficulty.
However, when it comes to modern complex,
scientifically demanding examinations requiring a deep
knowledge of chemistry, physics, or other sciences, for
example,
genetic
or
biological
examination,
physicochemical
research,
financial-economic,
computer-technical, and other similar forensic
examinations, assessing their results often presents
difficulties. Because for the purpose of assessing the
results of forensic examinations of this type and the
actions of the expert, the subject of assessment must
have minimum knowledge of the methodology of the
examination under consideration.
Currently, in the practice of law enforcement agencies,
such problems are solved as follows: preliminary
(preventive) verification by authorized bodies of the
correctness of the relevant research methods and
methodology and development of instructions for
forensic experts on the application of this methodology
(in cooperation with forensic experts); independent
education, advanced training, etc. of subjects of
assessment (for example, investigators, inquiry
officers, courts); obtaining explanations on the
necessary issues from the forensic expert who
conducted the examination, or from another specialist
in the relevant field.
After all, the forensic expert who conducted the
examination must be able to fully understand and
explain what actions he took during the examination
and exactly why he did so, what scientific principles his
actions were based on, where he could have made
mistakes, and why he obtained the specific results
noted in his conclusion.
However, the development of digital technologies and
the expansion of their use in forensic examinations
leads to the fact that it is no longer enough to be a
specialist in any field of knowledge, for example,
forensics, to conduct forensic examinations and assess
the accuracy of their results, and knowledge in the field
of computer technologies is also required. Initially, this
knowledge should have been at the level of an ordinary
user of computer technology, but with the expansion
of the scope of computer use, the required level of this
knowledge increases.
Today, in many forensic examinations, computer
technologies are used as an auxiliary tool in the
following processes: obtaining and processing digital
images, audio recordings, preliminary processing of
various databases located in computer memory,
computer modeling, comparison of digital copies of
objects, etc. In such cases, the forensic expert has the
opportunity to control computer operations and, if
necessary, verify the objectivity and validity of the
obtained results using other means.
With the advent of artificial intelligence, computers
became capable of independently performing a
number of separate (individual) practical intellectual
tasks, for example, recognizing a person's oral speech
or face (called "weak artificial intelligence") and even
creative thinking and self-awareness as a person
International Journal of Law And Criminology
27
https://theusajournals.com/index.php/ijlc
International Journal of Law And Criminology (ISSN: 2771-2214)
("strong artificial intelligence") [4]. This makes it very
difficult to effectively manage the "mental thinking
activity" of the computer.
In this regard, two questions arise regarding the use of
artificial
intelligence
for
conducting
forensic
examinations. One of them is that if the expert's
conclusion is incorrect, who will be responsible: the
forensic expert, the artificial intelligence manufacturer,
or the artificial intelligence itself? The second question
is related to the methodology for assessing the
accuracy of conclusions obtained as a result of expert
research. Since the first question is mainly related not
to the methodology of conducting the examination, but
to the legal regulation of the use of artificial intelligence
in criminal proceedings, let's focus on the second issue
in this case.
In this article, we did not examine cases where the
examination is conducted entirely by artificial
intelligence, and as a result, the forensic expert
receives a conclusion prepared only as a result of
research conducted by a computer system. Because in
this case, it is necessary to return to the first question -
who is the subject of the forensic examination.
In our research, we will focus on cases where the
subject of the examination is the forensic expert and,
accordingly, the forensic expert is responsible for the
accuracy of the obtained results. In this case, the
forensic expert must monitor the research process to
one degree or another and, if necessary, coordinate it.
When using artificial intelligence, this is a very difficult
task. Because in such a situation, the verification of
results is complicated by the multitude of reasons that
can lead to various errors. The following can lead to
incorrect results obtained during this activity using
artificial intelligence: errors in choosing a research
methodology; errors in the software of artificial
intelligence; errors in teaching artificial intelligence.
Errors in artificial intelligence training can be errors in
the selection of curricula or errors in the incorrect
selection of objects necessary for conducting artificial
intelligence training. Training on artificial intelligence
has a great influence on the objectivity of the results of
its use and application. Because this process is based on
the analysis of a large amount of data.
For artificial intelligence to function correctly, the data
used must be sufficient, systematized, and reflect all
possible situations. In training on artificial intelligence,
it is necessary to create independent models not only
for the initial state of the start of expert research, but
also for subsequent stages of research that determine
intermediate and final results, for example, situations
that reflect the degree of expression and the quantity
of features reflected in objects.
Some authors correctly emphasize that to solve each
new task of artificial intelligence, a large amount of
data is needed, based on specific and real examples
related to solving this problem, therefore, when
teaching artificial intelligence, it is necessary to take
into account all the tasks that artificial intelligence can
perform, and include in its memory examples of
solutions to these tasks [5, p.288]. For example, if
artificial intelligence doesn't have enough information
in its memory to solve any task, it can fill existing gaps
with what it considers right.
It should also be remembered that one of the functions
of artificial intelligence is the ability to learn
independently, in the process of which all new data
entering the system are analyzed for further use in
performing various tasks. Therefore, it is important to
regularly monitor the correctness of learning
algorithms embedded in artificial intelligence and the
appropriateness of newly introduced data.
Another factor that creates difficulties in assessing the
results obtained by artificial intelligence is the degree
of "closeness" of its work, the ability to study it
independently. The essence of this fact lies in the fact
that even the developers of such computer systems do
not have complete information about what happens
inside the program after the "independent life" of the
system begins, how the individual algorithms of its
operation change.
This creates a requirement for the forensic expert for
the transparency of artificial intelligence work (for
example, the possibility of constant monitoring and
supervision of processes). The forensic expert must
monitor the functioning of artificial intelligence at any
stage of the forensic examination, verify the
correctness of not only the final but also any
intermediate conclusion; the correctness of the system
of features selected for evaluation, etc.
Thus, it is possible to formulate the main requirements
for the processes of creating an artificial intelligence
system for conducting forensic examinations:
1. Artificial intelligence systems should be created with
the participation of forensic experts of the relevant
specialty, based on the methodology for conducting a
specific type of forensic examination. The processes of
developing, implementing, and using artificial
intelligence should be monitored by departmental and
interdepartmental organizations. This circumstance
protects developers from possible errors in the
selection of system algorithms and increases the
transparency of the system for forensic experts.
2. Artificial intelligence systems should be created on
the basis of transparency and accountability. The
subject responsible for conducting the forensic
International Journal of Law And Criminology
28
https://theusajournals.com/index.php/ijlc
International Journal of Law And Criminology (ISSN: 2771-2214)
examination must monitor and control the results of
artificial intelligence actions (operations) at each stage
of the forensic examination.
3. Training on artificial intelligence should be
conducted at specific objects under study, taking into
account all possible preliminary and intermediate
expert situations in the process of conducting certain
types of examinations.
4. Before implementing a specific artificial intelligence
system in expert practice, it is necessary to develop a
methodology for checking the correctness of its
operation and, if necessary, eliminating the errors that
have arisen.
5. In the process of using artificial intelligence, it is
necessary to periodically monitor the correct
functioning of artificial intelligence.
REFERENCES
Decree of the President of the Republic of Uzbekistan
dated October 5, 2020 No. UP-6079 "On Approving the
Strategy "Digital Uzbekistan - 2030" and Measures for
its Effective Implementation." / National Database of
Legislation,
06.10.2020,
No
06/20/6079/1349;
31.10.2020, No 06/20/6099/1450; 02.04.2021, No
06/21/6198/0269.
//
https://lex.uz/docs/5030957#5031883.
Criminal Procedure Code of the Republic of Uzbekistan.
187-m.
[Electronic
resource]:
-
URL:
https://lex.uz/docs/111460.
Farrux Hakimov. Sun’iy intellekt –
to‘rtinchi sanoat
inqilobining
asosi
/
https://strategy.uz/index.php?news=1198.
Маковкин А.С. Этические проблемы применения
искусственного
интеллекта
//
Издательство
«Грамота». 2015. № 2 (52). С. 130–
132.
Хмыз А.И. Экспертная ситуация и искусственный
интеллект // Вестник Московского университета
МВД России. 2022. № 3. –
С.286
-290.
