<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta />
    <article-meta>
      <title-group>
        <article-title>Cybersecurity Role in AI-Powered Digital Marketing</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <string-name>Vasyl Buhas</string-name>
          <xref ref-type="aff" rid="aff1">1</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Ihor Ponomarenko</string-name>
          <email>i.v.ponomarenko.stat@gmail.com</email>
          <xref ref-type="aff" rid="aff2">2</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Natalia Buhas</string-name>
          <xref ref-type="aff" rid="aff1">1</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Hennadii Hulak</string-name>
          <email>h.hulak@kubg.edu.ua</email>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <aff id="aff0">
          <label>0</label>
          <institution>Borys Grinchenko Kyiv Metropolitan University</institution>
          ,
          <addr-line>18/2 Bulvarno-Kudriavska str., Kyiv, 04053</addr-line>
          ,
          <country country="UA">Ukraine</country>
        </aff>
        <aff id="aff1">
          <label>1</label>
          <institution>Kyiv National University of Technologies and Design</institution>
          ,
          <addr-line>2 Mala Shyianovska str., Kyiv, 01011</addr-line>
          ,
          <country country="UA">Ukraine</country>
        </aff>
        <aff id="aff2">
          <label>2</label>
          <institution>State University of Trade and Economics</institution>
          ,
          <addr-line>19 Kyoto str., Kyiv, 02156</addr-line>
          ,
          <country country="UA">Ukraine</country>
        </aff>
      </contrib-group>
      <abstract>
        <p>Digitalization of human activity leads to a change in the types of economic activity and everyday behavior patterns of most consumers in the world. The development of innovative information technologies leads to the introduction to the market of qualitatively new products that are integrated into the Internet and are in high demand among modern users. The presence of a highly competitive environment encourages companies to identify ways of optimal development with the involvement of advanced approaches that will ensure high positions in the functioning market and form an economically justified demand for brand products from users. Achieving the presented tasks involves the use of modern marketing strategies in the digital environment. Thanks to the optimization of various digital marketing tools, it is possible to ensure long-term communications with the target audience and achieve a high level of loyalty among different groups of users. The use of modern web analytics tools allows generating information about various phenomena on the company's web resources. The development of server technologies makes it possible to accumulate large sets of heterogeneous data and process them thanks to the use of machine learning algorithms. The evolution of data processing methods, thanks to the use of various mathematical methods and models, has led to the widespread use of artificial intelligence in modern conditions. The effectiveness of artificial intelligence is explained by the ability to learn from large amounts of information and adapt the results of modeling to the changing influence of internal and external environmental factors. The above advantages led to the integration of artificial intelligence algorithms in digital marketing, which made it possible to ensure a qualitatively new level of digital tools implementation in the process of interaction with the target audience. Digital marketing communications led to the accumulation of various information about the company's activities and personal data of users, which involves the construction of a secure data storage system against the criminal actions of third parties. The approaches implemented in modern cybersecurity systems make it possible to reliably protect the information of all participants involved in marketing processes in the digital environment. The combination of digital marketing tools, artificial intelligence algorithms, and modern cybersecurity technologies allows to achieve a multiplicative effect in increasing the economic results of companies' activities.</p>
      </abstract>
      <kwd-group>
        <kwd>1 Artificial intelligence</kwd>
        <kwd>communications</kwd>
        <kwd>cybersecurity</kwd>
        <kwd>data</kwd>
        <kwd>digital marketing</kwd>
        <kwd>optimization</kwd>
        <kwd>target audience</kwd>
      </kwd-group>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>1. Introduction</title>
      <p>
        In the modern world, the concept of big data,
which is connected with the digitalization of
global and national economic systems, has
gained significant distribution [
        <xref ref-type="bibr" rid="ref4">1</xref>
        ]. Thanks to
powerful servers and specialized software,
companies have the opportunity to accumulate
information about various processes within
structural divisions. An important source of
data is the Internet, since a large number of
operations, including interaction with users,
are carried out by companies on the global
network. Information acts as a valuable
resource that must be analyzed using modern
machine learning algorithms and effective
management decisions are formed based on
the results obtained. Google Analytics 4 and
other web analytics tools allow companies to
accumulate large amounts of information
based on a customized system of metrics, and
the list of indicators may change due to
changes in the internal and external
environment of the company. Scientific and
technical progress and increased competition
stimulate the active development of digital
marketing tools that allow to accumulation of
new data and use the results of modeling based
on large volumes of information for further
qualitative improvement of strategies for
interaction with the target audience.
Opportunities to accumulate large volumes of
heterogeneous information in the process of
implementing marketing strategies in the
digital environment provide companies with
significant prospects for further development,
along with this a complex of threats arises
regarding the illegal use of data by third
parties. The use of information by criminals to
commit fraudulent activities can cause losses
to the company and its customers, as well as
lead to a deterioration of the image of the
respective brand. To minimize the risks of
illegal possession of information by criminals,
the company must implement an effective data
protection system. At the current stage of
development, there is a large number of
specialized cybersecurity systems that allow
companies to resist threats of theft of personal
data and information related to commercial
secrets [
        <xref ref-type="bibr" rid="ref5">2</xref>
        ]. Cybersecurity technologies involve
the use of a large number of algorithms, among
which the following must first be noted:
authentication and authorization,
cryptographic protocols, threat detection
methods, hash functions, digital signature,
encryption, etc. When implementing
information protection systems, it is expected
to find the optimal ratio between the cost of
digital security technologies and the obtained
economic effect [
        <xref ref-type="bibr" rid="ref6">3</xref>
        ]. Excessive spending of
monetary resources on the creation of a
company’s cybersecurity system may not
correspond to the value of the data that the
company owns, which will contribute to
significant economic losses [
        <xref ref-type="bibr" rid="ref7">4</xref>
        ]. Integrating
artificial intelligence into a company’s digital
marketing strategy allows companies to
identify hidden relationships in the data and
use the results to improve the company’s
performance. Along with this, it is possible to
connect artificial intelligence with the
company’s cybersecurity system, which will
make it possible to bring the data protection
system to a qualitatively new level [
        <xref ref-type="bibr" rid="ref8">5</xref>
        ].
      </p>
    </sec>
    <sec id="sec-2">
      <title>2. Related Works</title>
      <p>Information protection systems are constantly
being improved in connection with the use of
new technologies by criminals to acquire private
data. The presence of high demand for effective
cybersecurity systems prompts scientists in
different countries of the world to conduct
comprehensive research to identify more
effective approaches to the preservation of
various user groups’ data. The high technology of
modern cybersecurity is based on scientific
achievements in the field of information
technology and data protection at various levels
of processing and storage. Innovations in related
fields are integrated into protection systems,
which allows the identification of new directions
of cybersecurity development and optimization
of information flow processes at various levels of
company functioning.</p>
      <p>
        The work [
        <xref ref-type="bibr" rid="ref9">6</xref>
        ] reveals the essence of
cybersecurity and the main directions of its
implementation by companies in modern
conditions. Based on the surveys, the author
established the main strategies and
characteristic differences that influenced the
choice of the appropriate data protection
system in each of the studied companies. It has
been established that the preventive approach
is used by the majority of companies when
building cybersecurity systems, and the second
most popular is the protection methodology
using preventive actions.
      </p>
      <p>
        The work [
        <xref ref-type="bibr" rid="ref10">7</xref>
        ] is devoted to the combination
of cybersecurity and artificial intelligence. The
study revealed the features of generative
artificial intelligence integration in the
construction of effective cybersecurity
systems. The authors present strategies for
using artificial intelligence in the construction
of defensive and offensive information
protection strategies in modern conditions.
The potential risks that may arise in
connection with the use of ChatGPT by
attackers in the process of implementing
complex strategies for acquiring business
information and personal data of interested
persons have been identified.
      </p>
      <p>
        The Internet acts as an important source of
data for the company, and effective
management decisions are formed based on
the methods of intellectual information
analysis. In this aspect, it is important to pay
attention to the following article [
        <xref ref-type="bibr" rid="ref11">8</xref>
        ], which
examines the features of the intellectual
analysis application in cybersecurity. Thanks
to the application of modern methods of
researching large information arrays, effective
systems for auditing and detecting intrusions
into databases are being developed.
      </p>
      <p>Features of the use of cybersecurity
approaches in digital marketing are presented in
the work [9]. The authors proved the
effectiveness of implementing digital marketing
strategies as an innovative direction of economic
activity and the importance of ensuring the
inviolability of generated data for fraudulent
activities. Scientific research allows companies to
conclude the importance of integrating modern
cybersecurity technologies in the process of
combating cybercrime. Research is planned to
identify future data theft threats and develop
preventive measures.</p>
      <p>The work [10] reveals the features of
providing private data in the Metauniverse.
Scientists have proven the importance of a
comprehensive analysis of immersion
technologies, artificial intelligence, and
blockchain, which are used for the effective
functioning of the Metaverse. The use of
effective systems for the protection of private
information acts as an important prerequisite
for increasing the level of customer loyalty to
companies in the digital universe.
3. The Aim
The creation of an effective system for
protecting the company’s information at
various levels is one of the important strategic
tasks in today’s conditions. Communications
with users, partners, and competitors in the
digital environment led to the formation of large
volumes of specific information that can be used
to optimize the company’s strategy in the long
term. The existence of threats of illegal
acquisition of company information leads to the
need to create an effective cybersecurity system
that meets today’s requirements and can
quickly adapt to challenges and threats in future
periods. Thanks to a comprehensive analysis of
existing areas of information protection, it is
possible to identify approaches that can be
rationally used by the specifics of a particular
company’s activities and strategic goals, taking
into account the cost of cybersecurity
technologies and the economic efficiency of the
implemented measures. A significant increase
in the popularity of machine learning
algorithms and artificial intelligence use
involves conducting comprehensive research
on the vectors of information protection
technology development and obtaining a
multiplier effect due to the joint use of these
approaches. Identification of new methods of
data processing and technologies for involving
various objects in information exchange allows
for improving cybersecurity systems. Thanks to
the use of scientifically based approaches, the
probability of developing more effective
information protection systems that will
minimize the risks of acquiring and using data
for criminal purposes increases significantly.
Among the applied directions, it is advisable to
pay attention to digital marketing, because in
the process of interaction with users,
companies get access to a large amount of
personal data on legal grounds. Ensuring ethical
norms in marketing and maintaining a positive
image of companies involves the application of
advanced concepts in the field of cybersecurity.
It is appropriate to evaluate the role of artificial
intelligence in the implementation of marketing
strategies as an important element of ensuring
the integrity of the information space and
countering cyber threats. Modeling various
approaches to the construction of information
protection systems allows to identification of
optimal solutions and the building of a real
functioning system of countermeasures against
the illegal acquisition of business information
and personal data [11, 12].
4. Models and Methods
In the 21st century, cybersecurity systems have
undergone significant transformations due to
the intensive introduction of innovative
information technologies. The growing
demand for data protection systems in the
digital environment encourages the
continuous introduction of advanced
algorithms for processing large amounts of
information. Fig. 1 presents the main
cybersecurity algorithms that are used in
modern conditions in the construction of
effective systems for countering illegal
acquisition of private information.</p>
      <p>CYBERSECURITY ALGORITHMS
n
o
i
t
p
y
r
c
n
E
s
n
o
i
t
c
n
u
F
h
s
a
H
s
e
r
u
t
a
n
g
i
S
l
a
t
i
g
i
D
d
an n
n io
ito taz
a i
tic ro
tenhu tuhA
A
n
o
i
tce sd
t
teD teoh
ea M
r
h
T
s
l
o
c
o
t
o
r
P
c
i
h
p
a
r
g
o
t
p
y
r
C
y
t
i
r
and ecuS
rk re
o tu
w c
teN trsa
u
r
f
n
I
s
d
o
h
t
e
M
l
o
r
t
n
o
C
s
s
e
c
c
A
The effectiveness of cybersecurity algorithms
explains their active use in the implementation
of marketing strategies in the digital
environment. The features of the main
algorithms are disclosed below and the features
of their use in digital marketing are given.</p>
      <p>1. Encryption plays an important role in the
company’s data protection system, as it allows
transforming information into an unreadable
form for outsiders. The presence of special keys
for decryption allows access to primary data only
to authorized persons. The main types of
transformation of information into an
unreadable form are symmetric and asymmetric
encryption [15].</p>
      <p>Symmetric encryption involves the use of a
single key for data encryption and decryption.
The risk of the presented approach is the need
to provide access to the key to at least two
people, which leads to an increase in the
probability of third parties taking possession of
the decryption tool. Among the symmetric
encryption algorithms, the most widely used
include Advanced Encryption Standard (AES);
Blowfish; Camellia; Data Encryption Standard
(DES); International Data Encryption Algorithm
(IDEA); Serpent; and Twofish.</p>
      <p>Asymmetric encryption involves the use of
public and private keys. The public key is used
only to transform data into an unreadable form.
The private key allows decryption of the
transformed information by persons using the
specified tool. This approach minimizes the risks
of cybercriminals taking possession of the
private key and using the obtained data to
commit criminal acts. Asymmetric encryption
algorithms include Diffie-Hellman Key Exchange,
Digital Signature Algorithm (DSA), ElGamal,
Elliptic Curve Cryptography (ECC), Lattice-Based
Cryptography, Post-Quantum Cryptography
(PQC), Rivest-Shamir-Adleman (RSA).</p>
      <p>The use of modern cybersecurity
approaches to data encryption in digital
marketing allows companies to ensure the
confidentiality of business information and
personal data of users, the integrity of the
company’s information system, and helps to
achieve a high level of trust among all
participants. Payment tools, customer bank
details and transactions integrated into
marketing systems remain inaccessible to
outsiders thanks to encryption. Encrypting
emails minimizes the risk of personal data being
accessed for fraudulent activities. Content
plays an important role in the implementation
of marketing strategies in the digital
environment, which in certain cases must be
protected from unauthorized copying and
distribution.</p>
      <p>2. Hash functions allow to transformation of
large amounts of information into short hash
values with a specified length. The
development of information technologies
leads to the appearance of vulnerabilities in
existing hash functions and the development of
new, more secure hash functions that allow for
a high level of big data protection. At the
moment, the following hash functions are
among the most popular: Message Digest
Algorithm 5 (MD5), RACE Integrity Primitives
Evaluation Message Digest (RIPEMD), Secure
Hash Algorithm 1 (SHA-1), Secure Hash
Algorithm 256-bit (SHA-256), Secure Hash
Algorithm 3 (SHA-3), Whirlpool. The main
areas of hash functions used in cybersecurity
systems are:
• Data integrity check. Thanks to the use of
hash functions, it is possible to identify
the facts of data transformation in the
process of transmission. The end
consumer of the data has the opportunity
to determine the hash value of the
received data and conduct a comparative
analysis with the sent hash.
• Password protection. The use of hash
functions allows to storage of passwords
in a secure form since the external system
only contains a cache of a specific
password. When the user enters a
password, the hash for the specified
combination of access characters is
compared with the stored authentication
cache. When implementing this approach,
attackers do not have the opportunity to
access the real password, and it is almost
impossible to use a hash to generate an
access code with correct characters.
• Protection of digital signatures.</p>
      <p>Digitization processes have led to the
intensification of digital signatures used
as an effective tool for certifying the
authenticity of documents and messages.
Using hash functions makes it possible to
verify signed content by determining the
hash for a specific document and
comparing it using a public key to a
template digital signature.
• Protection on servers. In the process of
recording private information and keys
on specialized servers, it is advisable to
use hash functions for protection. In the
case of illegal access to the servers,
attackers will only be able to get hold of
the hash values of the passwords, which
have no value without decryption [16].
• Cryptographic protocols. This group of
approaches involves the use of
pseudorandom numbers and a system of
additional parameters that allow
generation hash functions with a high
level of protection. Thanks to the use of
cryptographic protocols, it is possible to
implement highly effective cybersecurity
systems for various participants in the
economic environment.</p>
      <p>The use of hash functions in digital
marketing makes it possible to achieve a high
level of information protection, including data
integrity, inviolability of personal information,
etc. The data obtained through the use of web
analytics tools can be depersonalized by
hashing methods. Information transformation
makes it impossible to leak personal data and
use it for fraudulent activities. Hash functions
allow access to photo and video materials only
to certain categories of consumers who have
received the appropriate permissions.</p>
      <p>3. Digital Signatures. The presented group
of algorithms makes it possible to identify the
originality of documents and establish cases of
information replacement by intruders. The
main algorithms include Digital Signature
Algorithm (DSA), Edwards-curve Digital
Signature Algorithm (EdDSA), Elliptic Curve
Digital Signature Algorithm (ECDSA),
Hashbased Message Authentication Code (HMAC),
Multifactor Digital Signature Algorithms,
Rivest-Shamir-Adleman (RSA).</p>
      <p>The companies’ implementation of complex
marketing strategies in the digital
environment includes the involvement of
innovative technologies, among which
machine learning algorithms, artificial
intelligence, and cybersecurity are of great
importance [17]. Fig. 2 shows the directions for
using digital signatures in digital marketing as
an effective cybersecurity tool for companies.</p>
      <p>To interact with certain categories of users,
it is advisable to use email marketing, which
involves sending themed emails with text
messages and specialized content. The use of
g
n
i
t
e
k
r
a
M
l
i
a
m
E
n
o
i
t
a
c
i
t
n
e
h
t
u
A
t
n
e
t
n
o
C
n
o
i
t
a
c
i
f
i
r
e
V
d
A
g
n
i
n
g
i
S
t
c
a
r
t
n
o
C
n
o
i
t
a
t
u
p
e
R
d
n
a
t
s
u
r
T
d
n
a
r
B
g
n
i
h
s
i
h
P
i
t
n
A
n
o
i
t
a
l
u
g
e
R
d
n
a
e
c
n
a
i
l
p
m
o
C
digital signatures allows for the transparency
of communications between companies and
users, making it impossible for cybercriminals
to falsify official e-mails of respective brands
during transmission.</p>
      <p>The growth of Internet users in the
conditions of digitalization stimulates the
active development of various types of
advertising. Thanks to the use of artificial
intelligence, Internet advertising is shown to
the target audience, which increases the
efficiency of interaction between companies
and users and allows for optimization of the
conversion rate. There are many players in the
online advertising market, which allows
cybercriminals to create fake advertising
messages to gain access to users’ confidential
information. Thanks to digital signatures, it is
possible to authenticate advertising messages
and identify criminal content. The use of digital
signatures in Internet advertising allows to
ensure the trust of advertising networks and
other market participants.</p>
      <sec id="sec-2-1">
        <title>DIGITAL SIGNATURES IN MARKETING</title>
      </sec>
      <sec id="sec-2-2">
        <title>INFORMATION SECURITY</title>
        <p>Today’s users, especially members of
Generation Z and Alpha, are interested in
learning new content regularly. To meet the
existing demand, companies must post
relevant photos, videos, audio, and text
information on social media, which helps to
retain the target audience. The creation of a
complete and secure information environment
for the company involves the provision of
content authentication. Users have the
opportunity to ensure their security in the
digital environment by checking the presence
of a digital signature on relevant content
associated with a specific company. This
approach is important because content
generation in today’s environment is becoming
more accessible thanks to the use of artificial
intelligence. Among the latest developments, it
is advisable to pay attention to the image
generation service based on the text
description, into which Dall-E 3 and ChatGPT
are integrated. The combination of Dall-E 3 and
ChatGPT allows to creation of images
containing the interaction of several complex
visualized objects. Accordingly, attackers can
create fake content that mimics the corporate
style of certain brands.</p>
        <p>Among the problems of modern digital
marketing, when building an effective
information protection system, it is necessary
to pay attention to the phishing of web
resources. Attackers impersonate legitimate
participants in the exchange of personal data
and illegally use passwords to access financial
and personal information. Fig. 3 shows the
main phishing methods.</p>
      </sec>
      <sec id="sec-2-3">
        <title>PHISHING METHODS</title>
        <p>ilac irng
oS ieen
g
n
E
4. Authentication and Authorization. In the
digital environment, authentication systems
based on machine learning algorithms, which
replace outdated methods based on entering
passwords and other symbolic information, are
becoming more widespread. Currently, the
following algorithms are used: Biometric
Authentication; Certificate-Based
Authentication; Kerberos; OAuth and OpenID
Connect; Multi-Factor Authentication (MFA);
Password-Based Authentication; Passwordless
Authentication; Public Key Infrastructure (PKI);
Single Sign-On (SSO); Smart Cards and PKCS#11;
Token-Based Authentication. Artificial
intelligence allows customers to use a person’s
face, voice, etc. as an identifier. The accuracy of
machine learning algorithms and the uniqueness
The next stage is authorization, which
represents the process of determining the level
of authenticated persons’ access to certain
information. The different level of access rights
to functionality and information is justified by
a complex of factors, including the options of a
certain digital product. For digital marketing
tools, authentication is a very important
feature, as it allows companies to create
products with special rates that reveal a clearly
defined set of capabilities to subscribers.</p>
        <p>5. Threat Detection Methods. The study of
large data arrays involves the use of various
and ssap
rs y
e B
t
il s</p>
        <p>u
F ir
am tiv
p n
S A
l
a
i
c
o
fseoS iaedM
U
of the physiological characteristics of each user
prevent unauthorized access to private
information by outsiders.</p>
        <p>
          The popularity of social networks among a
large number of users forces companies to create
complex strategies for interaction with the target
audience in these media. By socio-economic,
demographic, and psychological characteristics,
different groups of users permanently use
appropriate social networks for
communications. To ensure the protection of
accounts and personal data, social media users
can use two-step authentication (Fig. 4). The
presented approach involves the use of several
digital channels for confirming the owner of an
account on social media.
mathematical models to identify anomalous
cases that differ in distribution indicators from
the characteristics of the general population
[
          <xref ref-type="bibr" rid="ref14">21</xref>
          ]. To detect abnormal or criminal actions,
modern cybersecurity systems can use various
algorithms:
• Signature-Based Detection. The presented
method is implemented by comparing
signatures from the existing database and
performing operations. The high efficiency
of threat identification is due to the
presence of various digital threat
description templates in the databases,
but there are risks of inefficiency in the
case of threats with new technical
characteristics.
• Anomaly-Based Detection. Within a
certain system, a baseline of normal
behavior with statistically reliable levels of
deviation is established. Deviations
outside the limits are identified as
potential threats to the information
environment. This approach can be used
for known and new types of threats. The
complexity of configuring Anomaly-Based
Detection can lead to false identification of
threats.
• Heuristic Analysis. To identify threats,
rules, and algorithms are used, which are
adjusted to scientifically based standards
of behavior normality, which include
expert assessments, historical data, and
existing information system security
criteria. The flexibility of the presented
approach allows for the adaptation of the
threat identification system to specific
needs.
• Machine Learning and AI. The development
of mathematical algorithms for processing
big data and the appearance on the market
of powerful servers for processing
information contributed to the growing
popularity of machine learning and
artificial intelligence. The presence of
many machine learning algorithms makes
it possible to identify the optimal
approaches for creating effective
cybersecurity systems based on the
specifics of individual enterprises’
activities. The effectiveness of machine
learning and artificial intelligence used in
the field of cybersecurity is manifested in
the rapid response to adaptive and
complex cyber threats [
          <xref ref-type="bibr" rid="ref15">22</xref>
          ].
        </p>
        <p>In digital marketing, machine learning and
artificial intelligence algorithms have gained
significant popularity, as they allow companies
to increase the level of interaction with the
target audience and optimize economic results.
Along with this, the use of machine learning
and artificial intelligence in the cybersecurity
system of digital marketing allows achieving a
high level of information protection related to
commercial secrets and containing the
personal data of customers. Fig. 5 presents the
machine learning algorithms used to ensure
digital marketing information security.</p>
      </sec>
      <sec id="sec-2-4">
        <title>MACHINE LEARNING ALGORITHMS</title>
      </sec>
      <sec id="sec-2-5">
        <title>Random Forest</title>
      </sec>
      <sec id="sec-2-6">
        <title>Cluster Analysis</title>
      </sec>
      <sec id="sec-2-7">
        <title>Q-Learning</title>
      </sec>
      <sec id="sec-2-8">
        <title>SUPERVISED LEARNING</title>
      </sec>
      <sec id="sec-2-9">
        <title>UNSUPERVISED LEARNING</title>
      </sec>
      <sec id="sec-2-10">
        <title>Support Vector Machines</title>
      </sec>
      <sec id="sec-2-11">
        <title>Principal Component Analysis</title>
      </sec>
      <sec id="sec-2-12">
        <title>Deep Learning Models</title>
      </sec>
      <sec id="sec-2-13">
        <title>Isolation Forest</title>
      </sec>
      <sec id="sec-2-14">
        <title>REINFORCEMENT</title>
      </sec>
      <sec id="sec-2-15">
        <title>LEARNING</title>
      </sec>
      <sec id="sec-2-16">
        <title>Deep Deterministic Policy</title>
      </sec>
      <sec id="sec-2-17">
        <title>Gradients</title>
      </sec>
      <sec id="sec-2-18">
        <title>Multi-Agent Reinforcement</title>
      </sec>
      <sec id="sec-2-19">
        <title>Learning</title>
      </sec>
      <sec id="sec-2-20">
        <title>Ensemble Methods</title>
      </sec>
      <sec id="sec-2-21">
        <title>Logistic Regression</title>
      </sec>
      <sec id="sec-2-22">
        <title>Gradient Boosting</title>
      </sec>
      <sec id="sec-2-23">
        <title>Self-Organizing Maps</title>
      </sec>
      <sec id="sec-2-24">
        <title>Monte Carlo Tree Search</title>
      </sec>
      <sec id="sec-2-25">
        <title>Autoencoders</title>
      </sec>
      <sec id="sec-2-26">
        <title>Recurrent Neural Networks</title>
      </sec>
      <sec id="sec-2-27">
        <title>Gaussian Mixture Models</title>
      </sec>
      <sec id="sec-2-28">
        <title>Evolution Strategies</title>
        <p>The presence of many machine learning
algorithms allows us to test and create various
systems for protecting the information
environment of digital marketing. The main
directions of using mathematical algorithms in
this field are phishing detection, malware
detection, intrusion detection, user and entity
behavior analytics, ad fraud detection, sentiment
analysis, anomaly detection, dimensionality
reduction, network traffic analysis, content
recommendation, behavior analysis, content
clustering and tagging, botnet detection,
dynamic network security, optimizing
cybersecurity incident response strategies, user
access control, optimizing ad placement
strategies in digital marketing, recognize and
respond to evolving attack patterns, etc.</p>
        <p>6. Cryptographic Protocols. The presented
technology is used to protect information
transmitted over the Internet. Establishing
secure interaction between systems in the
digital environment is carried out through the
use of special key exchange protocols. Modern
cryptographic protocols include Hypertext
Transfer Protocol Secure (HTTPS), Internet
Protocol Security (IPsec), Open Authorization
(OAuth), Pretty Good Privacy (PGP), Secure
Sockets Layer/Transport Layer Security
(SSL/TLS), Secure/Multipurpose Internet Mail
Extensions (S/MIME).</p>
        <p>
          The use of cryptographic protocols in digital
marketing makes it possible to achieve a high
level of data confidentiality and authentication.
The presented technology is used during
interaction between users and web resources
when receiving information during the
implementation of advertising campaigns on
the Internet [
          <xref ref-type="bibr" rid="ref18">25</xref>
          ].
        </p>
        <p>
          7. Network and Infrastructure Security. This
group includes algorithms used to identify and
counter attacks in networks. The main
algorithms are Firewall Algorithms, Honeypot
and Deception Technology Algorithms, Intrusion
Detection and Prevention Algorithms (IDPS),
Intrusion Detection System (IDS)
SignatureBased Algorithms, Virtual Private Network
(VPN) Encryption Algorithms, etc. [
          <xref ref-type="bibr" rid="ref19">26</xref>
          ].
        </p>
        <p>8. Access Control Methods. The presented
methods enable different groups of users to
access different information by the established
rights. Modern algorithms include Adaptive
Access Control, Attribute-Based Encryption
(ABE), Biometric Access Control,
BlockchainBased Access Control, Multi-Factor</p>
        <p>Authentication (MFA), Role-Based Access
Control (RBAC), Single Sign-On (SSO),
Time-ofAccess Control, etc.</p>
      </sec>
    </sec>
    <sec id="sec-3">
      <title>5. Further Research</title>
      <p>The presented study reveals the specifics of
implementing effective cybersecurity systems
to protect information that is used to ensure
the functioning of digital marketing strategies
of various companies. Digitization processes
contribute to the intensification of legal and
illegal technologies development for
processing large arrays of heterogeneous
information. The integration of artificial
intelligence into cybersecurity systems makes
it possible to optimize the processes of
combating illegal acquisition of information for
fraudulent and criminal actions. Strengthening
the interaction between companies and users
in the digital environment through the use of
advanced marketing tools involves the
integration of advanced cybersecurity
approaches based on artificial intelligence
algorithms. Further scientific research
involves determining the optimal methods of
processing large amounts of marketing
information using machine learning methods
to ensure the protection of existing
information. It is appropriate to pay attention
to the features of the protection of visual
content of companies, as artificial intelligence
services for image creation are gaining
popularity in the market, for example, a
product with integrated Dall-E and ChatGPT.</p>
    </sec>
    <sec id="sec-4">
      <title>6. Conclusion</title>
      <p>Information protection is one of the priority
tasks in all types of economic activity. Building
an effective cyber protection system allows
companies to ensure the inviolability of
business data and personal information of
users. Observance of trade secrets allows
companies to gain advantages over
competitors in the market, and reliable storage
of customer information ensures a high level of
mutual trust and a loyal attitude of the target
audience. The emergence of new machine
learning algorithms used by artificial
intelligence stimulates the further
development of cybersecurity systems.
Artificial intelligence masters the elements of</p>
    </sec>
  </body>
  <back>
    <ref-list>
      <ref id="ref1">
        <mixed-citation>
          <article-title>creativity and can quickly respond to changes Mesopotamian J. Cybersecur. 2022 in internal and external environmental factors</article-title>
          , (
          <year>2022</year>
          )
          <fpage>28</fpage>
          -
          <lpage>37</lpage>
          . doi:
          <volume>10</volume>
          .58496/MJCS/ which will contribute to increasing
          <source>the 2022/004.</source>
        </mixed-citation>
      </ref>
      <ref id="ref2">
        <mixed-citation>
          <source>effectiveness of the protection of information</source>
          [9]
          <string-name>
            <given-names>S.</given-names>
            <surname>Kumar</surname>
          </string-name>
          ,
          <string-name>
            <given-names>H.</given-names>
            <surname>Pallathadka</surname>
          </string-name>
          ,
          <string-name>
            <surname>L.</surname>
          </string-name>
          <article-title>Pallathadka, systems used by companies in the process of An Analysis of Cybersecurity Threats in implementing marketing strategies in the Digital Marketing</article-title>
          ,
          <string-name>
            <given-names>J.</given-names>
            <surname>Crit</surname>
          </string-name>
          . Rev.
          <volume>9</volume>
          (
          <issue>03</issue>
          )
          <article-title>digital environment</article-title>
          . (
          <year>2022</year>
          )
          <fpage>85</fpage>
          -
          <lpage>94</lpage>
          . [10]
          <string-name>
            <given-names>M.</given-names>
            <surname>Pooyandeh</surname>
          </string-name>
          , K. Han,
          <string-name>
            <surname>I</surname>
          </string-name>
          . Sohn,
        </mixed-citation>
      </ref>
      <ref id="ref3">
        <mixed-citation>
          <article-title>References Cybersecurity in the AI-Based Metaverse: A Survey, Appl</article-title>
          . Sci.
          <volume>12</volume>
          (
          <issue>24</issue>
          )
        </mixed-citation>
      </ref>
      <ref id="ref4">
        <mixed-citation>
          [1]
          <string-name>
            <given-names>D.</given-names>
            <surname>Virovets</surname>
          </string-name>
          , et al.,
          <article-title>Ways of Interaction of (</article-title>
          <year>2022</year>
          )
          <article-title>12993</article-title>
          . doi:
          <volume>10</volume>
          .3390/ app122412993. Autonomous Economic Agents in [11]
          <string-name>
            <given-names>R.</given-names>
            <surname>Motoryn</surname>
          </string-name>
          , et al.,
          <article-title>Evaluation of Regional Decentralized Autonomous Organiza- Features of Electronic Commerce in tions</article-title>
          , in: Cybersecurity Providing in Europe, Stat.
          <source>J. IAOS</source>
          <volume>38</volume>
          (
          <issue>4</issue>
          ) (
          <year>2022</year>
          )
          <fpage>1339</fpage>
          - Information and Telecommunication 1347. doi:
          <volume>10</volume>
          .3233/sji-220938
          <source>. Systems</source>
          , vol.
          <volume>3421</volume>
          (
          <year>2023</year>
          )
          <fpage>182</fpage>
          -
          <lpage>190</lpage>
          . [12]
          <string-name>
            <given-names>I.</given-names>
            <surname>Gryshchenko</surname>
          </string-name>
          , et al.,
          <source>Making Use of</source>
        </mixed-citation>
      </ref>
      <ref id="ref5">
        <mixed-citation>
          [2]
          <string-name>
            <given-names>V.</given-names>
            <surname>Buhas</surname>
          </string-name>
          , et al.,
          <article-title>Using Machine Learning Competitive Advantages of a University Techniques to Increase the Effectiveness EEdduuccaattiioonnalISnenrovvicaetsioMnarCkleuts,tEeurroipneatnhJe</article-title>
          . of Cybersecurity, in: Workshop on Sustainable Dev.
          <volume>10</volume>
          (
          <issue>2</issue>
          ) (
          <year>2021</year>
          )
          <fpage>336</fpage>
          . Cybersecurity Providing in Information [13]
          <string-name>
            <given-names>C.</given-names>
            <surname>Annamalai</surname>
          </string-name>
          ,
          <source>Application of Factorial and Telecommunication Systems</source>
          , vol. and Binomial Identities in Information,
          <volume>3188</volume>
          , no.
          <issue>2</issue>
          (
          <year>2021</year>
          )
          <fpage>273</fpage>
          -
          <lpage>281</lpage>
          . Cybersecurity and Machine Learning,
        </mixed-citation>
      </ref>
      <ref id="ref6">
        <mixed-citation>
          [3]
          <string-name>
            <given-names>S.</given-names>
            <surname>Obushnyi</surname>
          </string-name>
          , et al.,
          <source>Ensuring Data Int. J. Adv. Netw. Appl</source>
          .
          <volume>14</volume>
          (
          <issue>1</issue>
          ) (
          <year>2022</year>
          )
          <article-title>Security in the Peer-to-</article-title>
          <string-name>
            <surname>Peer Economic</surname>
          </string-name>
          5258-
          <fpage>5260</fpage>
          . doi:
          <volume>10</volume>
          .35444/ijana.
          <year>2022</year>
          .
          <volume>14103</volume>
          .
          <article-title>System of the DAO</article-title>
          , in: Cybersecurity [14]
          <string-name>
            <given-names>M.</given-names>
            <surname>Sewak</surname>
          </string-name>
          ,
          <string-name>
            <given-names>S.</given-names>
            <surname>Sahay</surname>
          </string-name>
          ,
          <string-name>
            <given-names>H.</given-names>
            <surname>Rathore</surname>
          </string-name>
          ,
          <article-title>Deep Providing in Information and Telecom- Reinforcement Learning in the Advanced munication Systems II</article-title>
          , vol.
          <volume>3187</volume>
          (
          <issue>2021</issue>
          )
          <article-title>Cybersecurity Threat Detection and 284-292</article-title>
          . Protection, Inf.
          <source>Syst. Front</source>
          .
          <volume>25</volume>
          (
          <issue>2</issue>
          ) (
          <year>2023</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref7">
        <mixed-citation>
          [4]
          <string-name>
            <given-names>S.</given-names>
            <surname>Obushnyi</surname>
          </string-name>
          , et al.,
          <source>Autonomy of 589-611</source>
          . doi:
          <volume>10</volume>
          .1007/s10796-022
          <article-title>- Economic Agents in Peer-to-</article-title>
          <string-name>
            <surname>Peer</surname>
          </string-name>
          10333
          <article-title>-x</article-title>
          . Systems, in: Cybersecurity Providing in [15]
          <string-name>
            <given-names>R.</given-names>
            <surname>Sharma</surname>
          </string-name>
          ,
          <string-name>
            <given-names>S.</given-names>
            <surname>Dangi</surname>
          </string-name>
          ,
          <string-name>
            <given-names>P.</given-names>
            <surname>Mishra</surname>
          </string-name>
          ,
          <string-name>
            <surname>A Information</surname>
          </string-name>
          <source>and Telecommunication Comprehensive Review on Encryption Systems</source>
          , vol.
          <volume>3288</volume>
          (
          <year>2022</year>
          )
          <fpage>125</fpage>
          -
          <lpage>133</lpage>
          . Based Open Source Cybersecurity Tools,
        </mixed-citation>
      </ref>
      <ref id="ref8">
        <mixed-citation>
          [5]
          <string-name>
            <given-names>M.</given-names>
            <surname>Ansari</surname>
          </string-name>
          , et al.,
          <source>The Impact and 6th International Conference on Signal Limitations of Artificial Intelligence in Processing</source>
          , Computing and
          <string-name>
            <given-names>Control</given-names>
            <surname>Cybersecurity</surname>
          </string-name>
          :
          <string-name>
            <given-names>A Literature</given-names>
            <surname>Review</surname>
          </string-name>
          .
          <source>Int. (ISPCC)</source>
          (
          <year>2021</year>
          )
          <fpage>614</fpage>
          -
          <lpage>619</lpage>
          . doi: J.
          <source>Adv. Res. Comput. Commun. Eng</source>
          .
          <volume>11</volume>
          (
          <issue>9</issue>
          )
          <fpage>10</fpage>
          .1109/ISPCC53510.
          <year>2021</year>
          .
          <volume>9609369</volume>
          . (
          <year>2022</year>
          )
          <fpage>81</fpage>
          -
          <lpage>90</lpage>
          . doi:
          <volume>10</volume>
          .17148/ijarcce. [16]
          <string-name>
            <given-names>A.</given-names>
            <surname>Ahmed</surname>
          </string-name>
          ,
          <string-name>
            <given-names>W.</given-names>
            <surname>Ahmed</surname>
          </string-name>
          ,
          <source>An Effective</source>
          <year>2022</year>
          .
          <volume>11912</volume>
          . Multifactor Authentication Mechanism
        </mixed-citation>
      </ref>
      <ref id="ref9">
        <mixed-citation>
          [6]
          <string-name>
            <given-names>D.</given-names>
            <surname>Ghelani</surname>
          </string-name>
          , Cybersecurity,
          <source>Cyber Threats, Based on Combiners of Hash Function Implications and Future Perspectives: A Over Internet of Things, Sensors</source>
          <volume>19</volume>
          (
          <issue>17</issue>
          )
          <string-name>
            <surname>Review</surname>
          </string-name>
          , Authorea
          <string-name>
            <surname>Preprints</surname>
          </string-name>
          (
          <year>2022</year>
          ). doi: (
          <year>2019</year>
          )
          <article-title>3663</article-title>
          . doi:
          <volume>10</volume>
          .3390/s19173663.
          <fpage>10</fpage>
          .22541/au.166385207.73483369/v1. [17]
          <string-name>
            <given-names>A.</given-names>
            <surname>Ahmed</surname>
          </string-name>
          ,
          <string-name>
            <given-names>O.</given-names>
            <surname>Barukab</surname>
          </string-name>
          , Unforgeable
        </mixed-citation>
      </ref>
      <ref id="ref10">
        <mixed-citation>
          [7]
          <string-name>
            <given-names>M.</given-names>
            <surname>Gupta</surname>
          </string-name>
          , et al.,
          <article-title>From ChatGPT to Digital Signature Integrated into ThreatGPT: Impact of Generative AI in Lightweight Encryption Based on Cybersecurity and Privacy</article-title>
          ,
          <source>IEEE Access Effective ECDH for Cybersecurity</source>
          <volume>11</volume>
          (
          <year>2023</year>
          ). doi:
          <volume>10</volume>
          .1109/ACCESS.
          <year>2023</year>
          . Mechanism in Internet of Things,
          <volume>3300381</volume>
          .
          <source>Processes</source>
          <volume>10</volume>
          (
          <issue>12</issue>
          ) (
          <year>2022</year>
          )
          <article-title>2631</article-title>
          . doi:
        </mixed-citation>
      </ref>
      <ref id="ref11">
        <mixed-citation>
          [8]
          <string-name>
            <given-names>I.</given-names>
            <surname>Salem</surname>
          </string-name>
          , et al.,
          <source>Introduction to The Data</source>
          <volume>10</volume>
          .3390/pr10122631. Mining Techniques in Cybersecurity. [18]
          <string-name>
            <given-names>K.</given-names>
            <surname>Rahman</surname>
          </string-name>
          ,
          <article-title>Applications of Blockchain Technology for Digital Marketing: A systematic Review, Blockchain Technol</article-title>
          . Appl. Digi. Mark. (
          <year>2021</year>
          )
          <fpage>16</fpage>
          -
          <lpage>31</lpage>
          . doi:
          <volume>10</volume>
          .4018/978-1-
          <fpage>7998</fpage>
          -8081-3.
          <year>ch002</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref12">
        <mixed-citation>
          [19]
          <string-name>
            <given-names>E.</given-names>
            <surname>Gualberto</surname>
          </string-name>
          , et al.,
          <article-title>The Answer is in the Text: Multi-stage Methods for Phishing Detection Based on Feature Engineering</article-title>
          , IEEE Access 8
          <article-title>(</article-title>
          <year>2020</year>
          )
          <fpage>223529</fpage>
          -
          <lpage>223547</lpage>
          . doi:
          <volume>10</volume>
          .1109/ACCESS.
          <year>2020</year>
          .
          <volume>3043396</volume>
          .
        </mixed-citation>
      </ref>
      <ref id="ref13">
        <mixed-citation>
          [20]
          <string-name>
            <given-names>Zimbra</given-names>
            <surname>Two-factor Authentication</surname>
          </string-name>
          . URL: https://wiki.zimbra.com/wiki/Zimbra_ Two-factor_authentication
        </mixed-citation>
      </ref>
      <ref id="ref14">
        <mixed-citation>
          [21]
          <string-name>
            <given-names>T.</given-names>
            <surname>Chen</surname>
          </string-name>
          , et al.,
          <article-title>System-Level Data Management for Endpoint Advanced Persistent Threat Detection: Issues, Challenges and Trends, Comput</article-title>
          . Secur.
          <volume>135</volume>
          (
          <year>2023</year>
          )
          <article-title>103485</article-title>
          . doi:
          <volume>10</volume>
          .1016/j.cose.
          <year>2023</year>
          .
          <volume>103485</volume>
          .
        </mixed-citation>
      </ref>
      <ref id="ref15">
        <mixed-citation>
          [22]
          <string-name>
            <given-names>B.</given-names>
            <surname>Bebeshko</surname>
          </string-name>
          , et al.,
          <source>Application of Game Theory</source>
          ,
          <article-title>Fuzzy Logic and Neural Networks for Assessing Risks and Forecasting Rates of Digital Currency</article-title>
          ,
          <source>Journal of Theoretical and Applied Information Technology</source>
          <volume>100</volume>
          (
          <issue>24</issue>
          ) (
          <year>2022</year>
          )
          <fpage>7390</fpage>
          -
          <lpage>7404</lpage>
          .
        </mixed-citation>
      </ref>
      <ref id="ref16">
        <mixed-citation>
          [23]
          <string-name>
            <given-names>C.</given-names>
            <surname>Kumar</surname>
          </string-name>
          ,
          <string-name>
            <given-names>T.</given-names>
            <surname>Bharati</surname>
          </string-name>
          ,
          <string-name>
            <given-names>S.</given-names>
            <surname>Prakash</surname>
          </string-name>
          ,
          <article-title>Online Social Network Security: A comparative Review Using Machine Learning and Deep Learning, Neural Process</article-title>
          .
          <source>Lett</source>
          .
          <volume>53</volume>
          (
          <year>2021</year>
          )
          <fpage>843</fpage>
          -
          <lpage>861</lpage>
          . doi:
          <volume>10</volume>
          .1007/s11063- 020-10416-3.
        </mixed-citation>
      </ref>
      <ref id="ref17">
        <mixed-citation>
          [24]
          <string-name>
            <given-names>S.</given-names>
            <surname>Pinto</surname>
          </string-name>
          ,
          <string-name>
            <given-names>P.</given-names>
            <surname>Siano</surname>
          </string-name>
          ,
          <string-name>
            <given-names>M.</given-names>
            <surname>Parente</surname>
          </string-name>
          ,
          <article-title>Review of Cybersecurity Analysis in Smart Distribution Systems and Future Directions for Using Unsupervised Learning Methods for Cyber Detection</article-title>
          , Energ.
          <volume>16</volume>
          (
          <issue>4</issue>
          ) (
          <year>2023</year>
          )
          <article-title>1651</article-title>
          . doi:
          <volume>10</volume>
          .3390/en16041651.
        </mixed-citation>
      </ref>
      <ref id="ref18">
        <mixed-citation>
          [25]
          <string-name>
            <given-names>A.</given-names>
            <surname>Aly</surname>
          </string-name>
          , et al.,
          <article-title>Design of Symmetric-key Primitives for Advanced Cryptographic Protocols</article-title>
          ,
          <source>IACR Trans. Symmetric Cryptol</source>
          .
          <volume>3</volume>
          (
          <issue>2020</issue>
          )
          <fpage>1</fpage>
          -
          <lpage>45</lpage>
          . doi:
          <volume>10</volume>
          .13154/ tosc.v2020.
          <year>i3</year>
          .
          <fpage>1</fpage>
          -
          <lpage>45</lpage>
          .
        </mixed-citation>
      </ref>
      <ref id="ref19">
        <mixed-citation>
          [26]
          <string-name>
            <surname>T. AlMasri</surname>
            ,
            <given-names>M.</given-names>
          </string-name>
          <string-name>
            <surname>Snober</surname>
            ,
            <given-names>A.</given-names>
          </string-name>
          <string-name>
            <surname>Al-Haija</surname>
          </string-name>
          ,
          <article-title>IDPSSDN-ML: An Intrusion Detection and Prevention System Using SoftwareDefined Networks and Machine Learning</article-title>
          ,
          <source>1st International Conference on Smart Technology</source>
          , Applied Informatics, and
          <string-name>
            <surname>Engineering</surname>
          </string-name>
          (APICS) (
          <year>2022</year>
          )
          <fpage>133</fpage>
          -
          <lpage>137</lpage>
          . doi:
          <volume>10</volume>
          .1109/APICS56469.
          <year>2022</year>
          .
          <volume>9918804</volume>
          .
        </mixed-citation>
      </ref>
    </ref-list>
  </back>
</article>