Volume 05 Issue 06-2025
24
International Journal of Advance Scientific Research
(ISSN
–
2750-1396)
VOLUME
05
ISSUE
06
Pages:
24-35
OCLC
–
1368736135
A
BSTRACT
The author concentrates on the recommendations aimed at ensuring the reliability of digital evidence by
using (AI) artificial intelligence technologies. This article investigates the utilization of artificial intelligence
in the automation of the analysis of digital evidence. The authors explore contemporary methodologies
regarding the application of AI technologies for the processing, classification, and analysis of digital traces
within forensic practice. The paper evaluates the advantages associated with the integration of AI into the
processes of cybercrime investigation, which encompass the enhancement of processing speed for
extensive data sets, the identification of concealed patterns, and the automation of routine tasks. Particular
emphasis is given to machine learning, neural networks, and computer vision algorithms within the
framework of digital evidence analysis. Furthermore, the article addresses the challenges concerning the
Journal
Website:
http://sciencebring.co
m/index.php/ijasr
Copyright:
Original
content from this work
may be used under the
terms of the creative
commons
attributes
4.0 licence.
Research Article
Ensuring The Reliability Of Digital Evidences Through The
Application Of Artificial Intelligence (Ai)
Submission Date:
May 12,
2025,
Accepted Date:
May 30, 2025,
Published Date:
June 18, 2025
Crossref doi:
https://doi.org/10.37547/ijasr-05-06-04
Shomaxsudov Shoakrom Shomuratovich
Digital Forensics Research Institute of the Law Enforcement Academy of the Republic of Uzbekistan Head of
Digital Forensics Laboratory, Uzbekistan
Gallyamov Marsel Gabtulxayevich
Digital Forensics Research Institute of the Law Enforcement Academy of the Republic of Uzbekistan Digital
forensics laboratory Senior prosecutor-criminalist, Uzbekistan
Abdiaxatov Jakhongir Ravshan o’g’li
Digital Forensics Research Institute of the Law Enforcement Academy of the Republic of Uzbekistan Digital
forensics laboratory prosecutor-criminalist, Uzbekistan
Dadaboyeva Guzal Akbarjonovna
Teacher (ESP), Uzbekistan
Volume 05 Issue 06-2025
25
International Journal of Advance Scientific Research
(ISSN
–
2750-1396)
VOLUME
05
ISSUE
06
Pages:
24-35
OCLC
–
1368736135
reliability of results derived from AI systems, issues of legal regulation and ethical aspects of the use of
automated solutions in criminal proceedings.
K
EYWORDS
Digital evidence, artificial intelligence (AI) technologies, cybercrime.
I
NTRODUCTION
Digital evidence, encompassing items such as
electronic mail, SMS messages, and online
browsing records, is progressively regarded as
vital in legal inquiries and judicial proceedings. It is
imperative
to
manage
digital
evidence
meticulously to preserve its authenticity and
guarantee its acceptability within the judicial
system. So far, identifying and ensuring reliability
of digital evidence is not always simple meaning
sometimes it requires accuracy and carefulness
and time to comprehend. This article suggests
using AI (artificial intelligence) approach in
ensuring the dependability of digital evidences. By
exploring the interface between AI and digital
evidences, these studies promote cross-
disciplinary collaboration and innovation and help
to count on future characteristics and instructions
within the field. There are some questions to
discuss for understanding the topic deeply, thus
this article gives answer to the following questions.
HYPOTHESIS AND AIM
What is digital evidence? How it can be helpful to
maintaining forensic procedure? Digital evidence
refers to any electronic data stored or accessible,
including information from computers, mobile
devices and cloud services. It may include e-mails,
text messages, web browsing history, documents,
images, audio/video recordings and social media
content. Metadata (data on data such as time and
location) can also be valuable digital evidence.
2. The importance of digital evidence digital
evidence is used in a wide range of investigations,
from cybercrime to traditional crime. It can
provide crucial information on the location,
communication and intention of the suspect. It can
also help to reconstruct events, identify
perpetrators and build a strong case for
prosecution. However, there are little challenges to
regulate digital technologies in modern world.
1.Difficulties in Managing Digital Proof Data
Volatility: It's critical to rapidly safeguard digital
data because it can be easily lost or altered
Technological
Complexity:
Gathering
and
evaluating digital evidence can be difficult due to
the quick advancement of technology. Data
Overload: It may be challenging to locate pertinent
evidence due to the overwhelming amount of
digital data. Security and Privacy: One of the main
challenges is maintaining the integrity of digital
evidence
while upholding privacy. 4. Best Practices for
Managing the Preservation of Digital Evidence: To
Volume 05 Issue 06-2025
26
International Journal of Advance Scientific Research
(ISSN
–
2750-1396)
VOLUME
05
ISSUE
06
Pages:
24-35
OCLC
–
1368736135
guarantee that evidence is admissible in court, it
should be gathered and stored utilizing recognized
forensic techniques. Documentation: It is essential
to keep thorough records of the entire evidence
gathering and processing procedure. Chain of
Custody: Preserving a straightforward chain of
custody guarantees that the evidence hasn't
been altered. Use of Certified Tools: The
dependability of the evidence is increased by using
digital forensics tools that have been verified and
certified.Examples of Digital Evidence in Action
Cybercrimes: When it comes to the prosecution of
cybercrime such as fraud, phishing, and hacking,
digital evidence is essential.
Financial Crimes:Fraudulent financial transactions
can be found and the offenders can be identified
using digital proof. Conventional Crimes: In
traditional crimes, digital evidence can reveal
important details about the conversations,
whereabouts, and intentions of suspects.
What is happening to digital evidence in crime at a
global level?
According to the Cyber Security Incident
Management Guide in the private sector, there are
particular protocols that must be followed in order
to contain, investigate, and/or resolve cyber
security incidents (such as distributed denial of
service attacks, unauthorized access to systems, or
data breaches) (Cyber Security Coalition, 2015).
Recovering swiftly or gathering evidence are two
main approaches to managing a cyber-security
incident (Cyber Security Coalition, 2015): The first
strategy, "recover rapidly," focuses on containing
the incident to reduce damage rather than on data
collecting or preservation. Crucial evidence may be
lost due to its emphasis on quick response and
recovery. The second strategy keeps an eye on the
cyber security event and concentrates on digital
forensic tools to collect data and proof about the
incident. The recovery from the cyber security
issue is delayed due to its major focus on gathering
evidence. The commercial sector is not the only one
using these strategies. The private sector's strategy
differs depending on the organization and its aims.
Problem statement. Disruptive technologies like
artificial intelligence (AI), the Internet of Things,
drones, and crypto currencies, which can be
extremely dangerous instruments in the hands of
criminals, have made it possible for new crimes to
arise quickly. As a result, we constantly encounter
new types of digital crimes, hybrid crimes, crime-
as-a-service, cyber-attacks, political advocacy
campaigns, and cybercrime for war crimes, to
name a few. These crimes have a significant impact
on our society and, consequently, how we live our
lives. Therefore, in order to safeguard
organizations and individuals, our society needs a
significantly greater ability to detect and
investigate illegal conduct. Therefore, while
utilizing the potential of digital technology, it is
imperative that we deepen our understanding of
how these tools can be used against our society and
consistently make it more difficult for criminals to
use them successfully.
Volume 05 Issue 06-2025
27
International Journal of Advance Scientific Research
(ISSN
–
2750-1396)
VOLUME
05
ISSUE
06
Pages:
24-35
OCLC
–
1368736135
Fig.1. Handling of digital evidence
LITERATURE REVIEW
The role of AI in forensic science
Without a doubt, artificial intelligence plays a part
in forensic sciences, which suggests that we need to
learn more about its effects.For instance, the
European Forensic Science Area 2030 strategy
for2030 includes AI and future technologies , albeit
given the rate of digitalization now, 2030 seems a
little far off. Although certain ideas and
observations are discussed in this section, Geradts
and Franke, who provide state-of-the-art
implementations, offer a more thorough
analysis.Numerous instances demonstrate how
quickly technology is developing. The evolution of
airplanes from the first wooden models over a
century ago to the modern jets serves as an
example. Similar to this, forensic science has
evolved, and new techniques for analysis keep
opening doors to maximize both digital and
physical evidence of crimes. AI techniques come in
a variety of forms, with more to follow. AI-based
techniques, however, still fall short of what our
human brains are capable of. According to
Kahneman, the human brain can be divided into
two distinct systems: a quick, unconscious system
and a slower, conscious system that distributes
attention as needed. It is fair to assume that the AI
systems of today merely replicate the unconscious
portion of the human brain.1 Driving a car serves
as a basic illustration. Autonomous driving AI
techniques are "non-conscious." They are unable to
manage unforeseen circumstances for which they
are not prepared. Conversely, we have the ability to
activate our conscious system, process the novel
circumstance, and take action right away. However,
we may find ourselves using the non-conscious
brain function to drive to our old workplace rather
than to our new one if we are fatigued or distracted.
We must therefore be conscious of AI systems'
limits. AI systems outperform humans in other
domains, such as processing and retaining vast
amounts of diverse data that are beyond the
capacity of human brains. These days, AI
techniques outperform humans in several forensic
domains, such language and image recognition.
This suggests that man-machine interaction has to
be reviewed. We shouldn't undervalue the
1.identification
2.collection
3.acquisition
4.preservation
5.analysis & reporting
Volume 05 Issue 06-2025
28
International Journal of Advance Scientific Research
(ISSN
–
2750-1396)
VOLUME
05
ISSUE
06
Pages:
24-35
OCLC
–
1368736135
application of AI, for example, to handle the
massive volume of data. One example is the
difficulty of managing the 40 billion IP addresses in
the globe and their potential permutations while
searching the internet for criminal activity. We also
shouldn't overestimate AI because forensic
scientists will always be needed in the wake of
digital transformation, for example, to handle
reports and discoveries and employ AI-based
technologies. To improve the effectiveness and
precision of investigations, forensic science
optimization entails utilizing the advantages of
both
computer-based
and
human-based
methodologies.
In order to find matches in databases that can
identify a certain person or object, we analyze
traces, turn them into vectors of digital
information, and feed these strings
—
derived from
things like fingerprints and DNA
—
into different
classification algorithms. Using labeled datasets to
train algorithms to reliably classify data or
anticipate outcomes, predictive AI, such as
supervised learning, involves adding logic and
training the system to recognize the desired
outcome. By inverting the process and feeding the
system results instead, generative AI enables us to
produce new material based on the input we
provide. We will probably be inundated with
generative AI content related to crimes as well.
Generative AI is extensively used, for example, to
generate code, create graphics and videos from
text, and create big language models. Systems like
ChatGPT are trained to respond in-depth to our
prompts by following our directions. However, we
must understand that generative AI creates
knowledge based on the input we provide. This
suggests that we cannot completely rely on the
response to be honest or accurate if it has received
little training on a particular topic. The creation of
specialized "ChatGPTs" for usage in particular
applications is another issue we currently face. The
same is true for criminals, such as WormGPT,
which is made to help criminals with their
programming and hacking activities and enable
them to engage in destructive actions. Since the
cost of training is still too expensive, ChatGPT and
related technologies are being used extensively.
The dependability of software is questioned. For
example, how can we understand the error rate in
code produced by AI? The transition from AI apps
to AI assistants, which combine the capabilities of
generative AI in novel ways, is anticipated to be the
next big advancement. AI-enabled discussions are
one example, which combines a number of AI
approaches, including voice-activated big language
models, digital twins, and visualization tools. We
must pay attention to the ethical issues raised by
the application of AI. We must have faith that AI-
based systems are safe, deliver real outcomes, and
haven't had any of their parts compromised in
order to use them. We must comprehend and take
into account prejudice in AI systems just as we do
with traditional approaches because bias in all its
manifestations is significant. One example is bias in
training data, since AI systems are capable of
reproducing human prejudices. The majority of
training databases depict reality rather than the
ideal state of our society, such as in terms of gender
equality, explains ability, and justice. This suggests
that in order for us to trust AI systems as much as
we do more traditional tools and instruments, we
Volume 05 Issue 06-2025
29
International Journal of Advance Scientific Research
(ISSN
–
2750-1396)
VOLUME
05
ISSUE
06
Pages:
24-35
OCLC
–
1368736135
must provide training and education, as well as
conduct regular assessments and monitor AI
systems for biases. Furthermore, we must discuss
when computer-based techniques ought to be
verified based on the demographic characteristics
of the populations involved in the cases under
consideration, rather than a representative
distribution of these characteristics in the broader
community.
Fig.2.
Artificial Intelligence and the Society of the Future
https://www.cdh.cam.ac.uk/research/projects/ai-forensics/
Fig.3.
Proposed Digital Forensic Framework
Volume 05 Issue 06-2025
30
International Journal of Advance Scientific Research
(ISSN
–
2750-1396)
VOLUME
05
ISSUE
06
Pages:
24-35
OCLC
–
1368736135
1
Fig.4.
Cyber Crime Conviction Rate in India between 3013-15
Key features AI technologies in investigations
Numerous AI technologies are leading the way in
automated evidence processing, such as machine
transcription
and
translation,
automated
redaction, tracking persons of interest in videos,
and AI-driven metadata tagging. Below is an
explanation of how each contributes to
investigations:
Machine
transcription
and
translation: evidence is available in multiple
formats and may be in languages that are not
familiar to the investigation team. The processes of
transcription and translation can significantly
6.Rughani, P. H. (2017). ARTIFICIAL INTELLIGENCE BASED DIGITAL FORENSICS FRAMEWORK. International Journal of
Advanced Research in Computer Science, 8(8), 10–14. https://doi.org/10.26483/ijarcs.v8i8.4571
delay an investigation. Locating a service for these
tasks is challenging, and once found, obtaining the
results can take several days. Additionally, if a
substantial amount of evidence requires this
processing, the time frame can extend even further.
AI can greatly expedite this process, enabling
investigators to uncover evidence that may have
otherwise gone unnoticed.
Automated redaction: in numerous instances,
certain individuals may be present in evidence files
that require redaction. In the past, this necessitated
a labor-intensive manual process that diverted
Volume 05 Issue 06-2025
31
International Journal of Advance Scientific Research
(ISSN
–
2750-1396)
VOLUME
05
ISSUE
06
Pages:
24-35
OCLC
–
1368736135
valuable resources from other critical tasks. AI can
significantly streamline the redaction process,
substantially decreasing the time required.
Person-of-interest and vehicle tracking: facial
recognition technology has limitations in scenarios
where video footage includes large crowds. While
there are appropriate contexts for its use, some
individuals express hesitation due to concerns
regarding the safeguarding of personally
identifiable information (PII). However, recent
advancements in AI technology that categorize
humans as objects can assist investigators in
identifying key features of a potential suspect (such
as a logo on a shirt, a specific type of hat or
backpack they are wearing, etc.) to monitor that
individual across video files on a large scale,
thereby greatly minimizing the time needed to
analyze video evidence. This same technology can
also be employed by investigators to track vehicles
within the footage.
AI-driven metadata tagging: when analyzed using
AI, digital evidence generates a substantial amount
of data that can be utilized to assist investigators in
comprehending the materials at their disposal.
From highlighting only pertinent files to identifying
moments within files that may contain crucial
evidence capable of altering the case, AI offers
automated evidence processing to equip
investigators with the insights necessary to locate
that needle in a haystack.
Ethical considerations in the responsible
deployment of AI As artificial intelligence
increasingly influences investigations and law
enforcement, it is essential that ethical
considerations
remain
paramount
in
its
implementation. Concerns such as algorithmic
bias, data privacy, and other related issues
necessitate that AI leaders adhere to standards
designed to safeguard individuals from potential
harm. While legislation is still striving to keep pace,
there are proactive measures that companies can
undertake immediately. Veritone has established a
framework of AI for Good principles that directs all
our efforts to ensure that our technology is
transparent, reliable, secure, and compliant.
Furthermore, it enhances the capabilities of
individuals, including investigators, enabling them
to perform their duties more effectively.
Research methodology introduction
This is the main part of the research work that
shows the procedures and methods used in this
study. The chapter is organized into subheadings
representing the steps of the study. These sub-
headings include: the digital forensic material to be
studied, research design, research instruments,
methods that are used to collect data and methods
which are used to analyze data.
Research design
Qualitative research method used in the form of
graphs, and charts which includes well-structured
some questions which show strategies are
necessarily give answers and 1 of them was open
question where participants are free to answer.
Population of the study
This study is based on secondary data materials,
basically taken from different articles and web
Volume 05 Issue 06-2025
32
International Journal of Advance Scientific Research
(ISSN
–
2750-1396)
VOLUME
05
ISSUE
06
Pages:
24-35
OCLC
–
1368736135
pages written by international scholars and
scientists.
Method of data analysis
Descriptive statistics were used to understand the
preference of the criminals on application of AI in
analyzing digital evidence.
Hypothesis 1
The role of digital evidence and its’ importance to
maintaining forensic procedure.
Digital evidence refers to electronic data used in
legal proceedings. It includes emails, texts and
more. Modern law enforcement relies heavily on
digital evidence to solve crimes and ensure
convictions. This article deals with the collection,
legal considerations and investigative role of
digital evidence. Digital evidence, which includes
data from various electronic sources, is crucial to
modern law enforcement, often surpassing DNA
evidence in the context of investigation. Special
methods and legal considerations are crucial to the
collection and analysis of digital evidence in order
to ensure its integrity and its admissibility to court,
including the maintenance of a strict chain of
custody. Technological advances, such as cloud
computing and artificial intelligence, have had a
major impact on digital law enforcement, requiring
constant updates to the tools and methodologies
used by investigators. Digital evidence refers to all
electronic data or information that can be used in
legal cases. This includes a wide range of digital
data, including: logging emails multimedia file
databases various forms of software record Digital
evidence plays a central role in investigations,
providing insight essential to the resolution of
crimes and the assurance of convictions.
Hypothesis 2
Utilizing the potential of AI technologies in digital
evidence analysis, and how these tools can be used
against our society and consistently make it more
difficult for criminals to use them successfully.
One of the primary benefits of artificial intelligence
in the management of digital evidence is its
capability to swiftly and accurately process
extensive amounts of data. Investigations
frequently require the examination of a vast array
of digital evidence from multiple sources, including
social media, surveillance videos, and emails. The
ability of AI to analyze these files and highlight
pertinent data enables investigators to concentrate
on the most essential aspects of the case without
being overwhelmed by manual data review tasks.
By employing machine learning algorithms and
natural language processing (NLP), AI tools can
quickly extract relevant information, assist teams
in recognizing patterns, and deliver crucial
insights. By implementing this approach, AI
introduces an additional level of accountability
aimed at minimizing human error, which tends to
occur more frequently in manual evidence analysis
procedures. AI addresses this risk by automating
repetitive tasks and examining evidence with
unwavering precision. Machine learning models
that have been trained on extensive datasets can
reveal and detect elements that might be
overlooked by human analysts. This improvement
boosts the accuracy of data analysis and
accelerates the investigation process, ultimately
Volume 05 Issue 06-2025
33
International Journal of Advance Scientific Research
(ISSN
–
2750-1396)
VOLUME
05
ISSUE
06
Pages:
24-35
OCLC
–
1368736135
resulting in more dependable and favorable
outcomes. So that, this article helps to identify key
features of using AI approach in two ways first for
its quickness for finding materials and second for
the effectiveness of AI approach using.
DATA INTERPRETATION AND FINDINGS
How AI approach may foster the work of
investigators?
1.Case In India, financial crime has developed to fit
a complicated ecosystem. For example, several
payment banks may encounter regulatory
consequences after the Financial Intelligence Unit
(FIU) found that nearly 50,000 accounts do not
have adequate Know Your Customer (KYC)
verification. These accounts are believed to be
linked to dubious transactions and possible
money-laundering activities. Of these accounts,
approximately 30,000 are associated with payment
banks excluding Paytm Payments Bank. Over the
last ten years, the Enforcement Directorate (ED)
documented its peak number of money laundering
and foreign exchange violation cases in 2021 and
2022, with 1,180 and 5,313 complaints,
respectively. Between FY 2012-13 and 2021-22,
the agency filed a cumulative total of 3,985 criminal
complaints under the Prevention of Money
Laundering Act (PMLA) and 24,893 under the civil
law of the Foreign Exchange Management Act
(FEMA). In the previous three years, the ED has
received more than 12,000 complaints regarding
alleged foreign exchange violations. Furthermore,
the adjudicating authority of the PMLA verified
proceeds of laundering totaling INR 2,214.92 crore
over the last three years.
Fig.5. Money Laundering and Foreign Exchange Management Act (FEMA) Cases in India
Volume 05 Issue 06-2025
34
International Journal of Advance Scientific Research
(ISSN
–
2750-1396)
VOLUME
05
ISSUE
06
Pages:
24-35
OCLC
–
1368736135
As illustrated in Fig.6 below, supervised learning is
the most commonly employed method in crime
prediction, accounting for 31% of the research
papers. Additionally, since some studies
implemented
multiple
machine
learning
algorithms, 22% of the collected papers utilized
both supervised and unsupervised methods. In
contrast, only 10% of the papers focused on
unsupervised learning. Interestingly, a mere 1%
adopted the semi supervised approach, indicating
that this method is rarely used in the realm of crime
prediction. Lastly, 36% of the papers did not clarify
which approach they adopted.
Fig.6. Learning methods percentages
C
ONCLUSION
This research underscores the transformative
potential of Artificial Intelligence (AI) in enhancing
the reliability and efficiency of digital evidence
analysis for forensic investigations. Key findings
demonstrate that AI technologies
—
including
machine learning, neural networks, computer
vision, and natural language processing
—
significantly accelerate the processing of large-
scale digital datasets, automate labor-intensive
tasks (e.g., transcription, redaction, metadata
tagging),
and
uncover
hidden
patterns
imperceptible to human analysts. These
capabilities are critical in combating evolving
cybercrimes, financial fraud, and hybrid threats.
R
EFERENCES
Volume 05 Issue 06-2025
35
International Journal of Advance Scientific Research
(ISSN
–
2750-1396)
VOLUME
05
ISSUE
06
Pages:
24-35
OCLC
–
1368736135
1.
AI IN DIGITAL FORENSICS Manasi Pritam
Zirpe1, Shravani Santosh Potdar1&,
Harshali
Rohit
Kadaskar.
(n.d.).
International Journal of Scientific Research
in Modern Science and Technology.
2.
Artificial intelligence & crime prediction: A
systematic literature review. (n.d.). Social
Sciences & Humanities Open Volume 6,
Issue 1, 2022, 100342.
3.
Muhammad Arjamand1, Areeba Saleem1 ,
Abdul Basit1 , Subha Iftikhar1 , Muhammad
Sharif2 , Muhammad Shahid Cholistani1 ,
Muhammad Farhan1 , Shumail1 , Bakht
Ameer Khan1 , Zeeshan Ali1 , Bilawal
Shahid1 , Muhammad Hasnain1. (n.d.).
International Journal of Multidisciplinary
Research and Publications.
4.
Mitchell, F. (2014). The use of Artificial
Intelligence in digital forensics: An
introduction.
Digital
Evidence
and
Electronic Signature Law Review, 7(0).
https://doi.org/10.14296/deeslr.v7i0.192
2
5.
Klasén, L., Fock, N., & Forchheimer, R.
(2024). The invisible evidence: Digital
forensics as key to solving crimes in the
digital age. Forensic Science International,
362,
112133.
https://doi.org/10.1016/j.forsciint.2024.1
12133
6.
The role of AI in forensic science. (n.d.).
Wiley.
7.
Rughani, P. H. (2017). ARTIFICIAL
INTELLIGENCE
BASED
DIGITAL
FORENSICS FRAMEWORK. International
Journal of Advanced Research in Computer
Science,
8(8),
10
–
14.
https://doi.org/10.26483/ijarcs.v8i8.4571
8.
SAURADEEP BAG, Use of AI in Arresting
Financial Crime. (n.d.). Published on Aug 19,
2024.
