RESEARCH METHODOLOGY IN COMPUTING &

TECHNOLOGY

Artificial Intelligence to detect and secure security threats

CT098-3-2-RMCT

Name :Huzaifa Bin Awais

TP Number : TP053594

Intake Code : APD2F2011CS(CYB)

Submission date : 9th May 2022

Lecturer’s name : Dr. Siti Sarah Binti Maidin

 Table of Contents
1. Introduction. 3

2. Research Background. 3

3. Problem Statement 4

4. Aims and Objectives. 5

5. Research Questions. 5

7. Methodology. 6

8. Overview of the Proposed System... 6

9. Conclusion. 6

10. References.11

11. Appendices. 11

Artificial Intelligence to detect and prevent security threats

Huzaifa Bin Awais
[email protected]
Abstract
The paper aims to debate regarding the Introduction
impact of cyber-attacks on organizations Cyber attacks have become more
and how A.I can be implemented to common around the world, posing a
detect and prevent attacks.Several threat to business operations. The
studies on safety threats and advancements in information technology
countermeasures have been conducted (IT) allow everything to be
over the years. The findings indicate an interconnected. Today, the vast majority
increase in safety threats as the use of of people and all gadgets/devices are
information technology grows rapidly, linked in cyberspace. Advances and
making it vulnerable to threats and trends continue to emerge, raising
intrusion. When we consider the critical questions about information
massive amounts of quantity statistics security. Because of the excessive use
that are used, as well as the surge in of the internet, it is more vulnerable to
cyber interest, we want intelligent security threats. As a result, we need
systems. It cannot be treated with the intelligent solutions, such as Artificial
aid of physical devices, and human Intelligence (AI), to combat security
intervention is clearly no longer threats. Because of its ability to think
sufficient. We want to discover modern like a human and make rational
strategies, such as the use of Artificial decisions, AI opens the door to many
Intelligence strategies, that can be used solutions. Furthermore, the fact that
to create intelligent systems capable of cyber criminals are using AI-powered
detecting and responding to security threats to assault/attack the IT
threats. infrastructure makes it even more
An in-depth literature review was vulnerable. Such AI systems are
conducted during this research paper to required to improve existing detection
demonstrate how cyber-crime affects rates and response capabilities. This
the day-to-day operations of company study examines how artificial
entities. During this analysis, intelligence (AI) can help prevent and
quantitative data collection methodology detect security threats, as well as why AI
was used to collect data from credible is important to us. Following that, there
secondary sources and frame will be a review of previous research on
constructive conclusions about the AI and security threats. It then
devastating impact of a cyber-attack on addresses the problem statements that
an organization's daily operations and have arisen from previous research, and
how A.I can help. it continues with the aim and objectives
(Artificial Intelligence, Threats, Cyber of this research, as well as what the
Attacks, and Security are all index study's ultimate goal is. This is followed
terms.) by a statement about the research
questions that this study intends to 5% of company information is
address. Following that, the significance successfully and efficiently secured.As a
of this research is discussed, followed result, artificial intelligence is extremely
by Methodology. The proposed important because it reduces human
strategies for gathering facts are errors, provides a more efficient and
discussed in methodology. Lastly, the effective method of detecting and
study ends with an outline of the preventing cyber attacks, and reduces
proposed device which addresses the data breaches in large corporations.
modern-day problems. Statistics/information breaches totaled
36 billion during the first half of 2020.
The majority of breaches and attacks
are carried out for monetary gain [8].
Companies with remote employees
must stabilize their statistics because
not securing documents increases the
opportunities and risks of a cyber
attack. Cybercriminals must face severe
AI in Cyber-Security penalties because the law is strict in the
case of cybercrime around the world..
(McKenna, B., 2018.)

Research background

The businesses are shielding

themselves from cyberattacks or threats
because of an upward thrust in crimes Research Aim
throughout the globeGlobally, the
statistics protection market is expected The goal of the research is to use
to grow to 170.4 billion dollars by 2022. artificial intelligence to detect and
Human error has resulted in security
prevent cyber attacks, as well as to
breaches throughout the business
determine and examine the impact
world. Human error is responsible for
approximately 95% of security breaches. of criminal attacks on commercial
Throughout 2019, approximately 88 enterprise organizations. As the
percent of business organizations are number of attacks increases
familiar with "spear phishing." Globally, globally, it is critical to determine
approximately 68 percent of high-level the impact of cyber-attacks on
business organizations have felt the commercial enterprise
threat of cyber-attacks. Approximately organizations.
 2. Human intervention or any bodily tool
is truly now no longer enough.
Conventional-programming algorithms
used aren't effective.
Research objectives

Several goals are developed in order to The Research questions are as follows:
achieve the research goal.To evaluate -
current literature on the use of Artificial
Intelligence to overcome cybercrime. 1. What is the best way to overcome
such attacks with AI?
● To learn about the skills and
opportunities that Artificial 2. What strategies are available for
Intelligence offers. updating traditional algorithms and
● Because of its adaptability and implementing AI?
learning abilities, Artificial
Intelligence can be used to create 3. Can AI enhance security and reduce
an intelligent system. cyberattacks further?
● To understand how attackers use
Artificial Intelligence for The following objectives are set with the
malicious purposes. intention of answering the questions
● To investigate how threats such raised above.
as malware behave.
● Finally, support a system that can 1. To learn about the possibilities of
detect and prevent security Artificial Intelligence and how it can be
threats with a high detection rate, used.
as well as respond to attacks 2. To select cutting-edge Artificial
quickly and effectively based on Intelligence strategies that can examine
the severity of the attack. and respond quickly, as well as assist
people involved in cybercrime.

Research questions
Problem statement
Research questions are derived from
two research problems which are: A minimum of 1,000,000 computer
systems are affected each year as a
1. With the ever-growing utilization of IT result of [7] computer virus attacks.
and cyber infrastructure they're at risk of These attacks are estimated to be worth
attacks. billions of dollars each year. This could
be due to the use of traditional
programming algorithms in structures Artificial intelligence's power is
that have been described as ineffective increasing exponentially. As Artificial
in preventing cybercrime via [3]. Intelligence becomes more powerful, we
According to [8], with the ever-increasing will use it in the field of cyber-security to
use of the Internet, network, and cyber detect security threats such as malware,
infrastructure, they are vulnerable to etc. [1] compiles a number of definitions
attacks, so there is a need for intelligent of Artificial Intelligence from notable
systems to help defend against cyber- authors. According to one author, it is
attacks, and because everything these the science of automating intelligent
days is interconnected, viruses can behaviours. According to the other
spread rapidly. As a result, intelligent author, it is intelligent machines that can
systems are required to combat cyber- self-learn, reason, and communicate.
assaults, which could prevent or And, according to the third author
mitigate an attack quickly and efficiently mentioned by [1], it is far defined as
without the need for human demonstrated intelligence that can be
involvement. used to solve complex problems.
Based on the definitions above, we can
This is supported by [2], which states conclude that artificial intelligence
that due to the increase in cyber-crime refers to intelligent systems that can
activities, human intervention is no self-master and make real-time
longer sufficient. decisions.
Malware and viruses that have been
released recently have a mechanism
that allows them to self-regulate and Applications to protect a company's
change their signature in order to avoid cyber infrastructure can be developed
detection. As a result, it's far stated in using the most recent artificial
this research that it's far critical to intelligence technologies.Cyber
examine and study the execution of infrastructures, according to [2,] are
those malware, as well as classify them. extremely vulnerable to threats. As a
This can lead to a greater understanding result, the advancements made thus far
of how such threats operate with the in the use of artificial intelligence
assistance of Artificial Intelligence. techniques to combat cybercrime are
discussed. Such strategies are capable
of detecting and preventing cyber-
attacks. The article also discusses
various methods of computing
Literature Review
intelligence, such as machine learning,
data analysis, artificial neural networks,
As organisations and businesses grow
pattern processing, and so on. Because
in size, there are new ways to deal with
of the increase in cyber activity and the
security that are becoming popular.
large amount of data used, [2] claims by using [4], because the authors state
that human intervention or any physical that it is difficult to develop/create any
device is insufficient for analyzing and software programme with traditional
responding to an attack.This claim is algorithms to defend against attacks.
supported by similar research
conducted by others; [3] and [4] also As a result, there may be a need for
claim that physical devices and human revolutionary artificial intelligence
intervention are no longer effective in strategies that can examine and
protecting the Information Technology respond to cybercrime instantly.
infrastructure due to the increase in
cyber-crime interest. In addition, [3] investigates the
application of artificial intelligence
As a result, numerous studies are being possibilities such as artificial neural
conducted in order to implement networks, deep learning, computational
intelligent systems. [5] is a follow-up to intelligence, machine learning, pattern
ENISA's research (European Network recognition, and so on. Whereas [4]
and Information Security Agency). The proposes a novel approach, such as an
findings had been on fundamental artificial intelligence-powered system
characteristics; certainly considered one capable of learning and assisting
of those to combine artificial intelligence humans in cybercrime.
techniques and technology to neutralize Another study [10] proposes a
cyber-attacks. framework for implementing machine
The corporation explained two research learning algorithms that can
conclusions: one concluded the successfully differentiate between
combination of artificial intelligence with malware and secure files.
cyber threat intelligence, which also Traditional defense systems, according
includes attack pattern recognition and to [9], use signature-based techniques
knowledge discovery. The other study that are incapable of detecting malware.
concluded the development of security To address this limitation, the authors
models using cyber intelligence to propose a malware analysis technique
evaluate the efficiency and performance that can help in the analysis of
of security controls. The research data malicious code in malware as well as
can be used to combine the two trends the understanding of its risks and true
and create a system with an algorithm intentions.
that can detect attacks. This review discusses all of the current
issues in the information technology
Conventional programming algorithms, infrastructure, as well as proposed
as stated with the help of [3], are solutions. Several studies have
ineffective in combating cybercrime. demonstrated artificial intelligence's
This assertion can indeed be supported potential and capabilities in assisting us
in defending against cyber-attacks. of strategies or algorithms can be used
Artificial intelligence assists in to update traditional algorithms?
addressing these issues due to its This study will use a descriptive
intelligent nature and flexibility. approach to answer the questions
Simultaneously, in order to improve raised above.
prevention and mitigation efforts, we This technique is used because this
must understand the malicious use of research describes cutting-edge issues
artificial intelligence. in security as well as trends in artificial
intelligence. The goal is to familiarize
the subject and benefit from a higher
level of expertise in the subject. This
research would necessitate the
collection of primary and secondary
data in order to achieve the research
objectives. The primary statistics
METHODOLOGY gathered will be directly related to the
purpose of this research, whereas
For this research, quantitative survey secondary data is second-hand
study will be used to determine the information gathered from other
possible solutions needed for the two sources and is unrelated to the purpose
research questions Quantitative of this research. Questionnaires
research establishes good sized distributed to enterprise companies and
conclusions about a populace through subject-matter experts may be used to
reading the facts collected [11].By collect primary data.To collect primary
utilising this method, it will be possible data from questionnaires, a random
to collect statistics on what types of sampling method may be chosen. This
strategies can be used to combat such sampling method allows for the
attacks. There are existing systems for collection of records from a subset of
such attacks that use standard the population with a pattern length of
programming algorithms that are known 200 to 300. A questionnaire designed to
to be ineffective, which leads to the collect data can include both closed-
other research question of what ended and open-ended questions.
techniques can be used to replace Closed-ended questions have a limited
conventional programming algorithms. set of alternatives, whereas open-ended
This survey method will use a sample of questions must be answered
300 people to gather data in order to descriptively. Having only closed-ended
assist in studying the strategies that are questions may result in biases, so it is
currently available in the field of Artificial recommended that open-ended
Intelligence. Following that, what kinds questions be included as well. The
questionnaire will include questions
about Artificial Intelligence trends and prevent Security threats.Currently, many
how traditional algorithms can be detection systems rely on malware
improved.After at least 150 responses signatures that are stored in a database.
have been received, the data collected This has been proven ineffective
will be processed, analyzed and because malware is capable of self-
documented. modifying and changing their signature
in order to avoid detection. As a result,
this study proposes an intelligent
machine that does not rely on
Data Analysis signatures and employs Artificial
Intelligence strategies to detect and
Several tools are used to interpret and prevent all types of threats while
present the facts derived from the maintaining a high detection rate. This
collected data. The research responses system should be able to self-learn in
can be presented using Microsoft Excel order to comprehend how threats such
to provide an overview of the responses. as malware execute and behave. The
Microsoft Excel was chosen because it machine must be capable of making
could be used for statistical analysis. rational decisions in a timely and
Excel has a number of tools that can be effective manner in response to an
used to collect descriptive data, attack.In the event of an incident, the
including data analysis ToolPak and system can protect you from dangerous
Pivot Tables. Pivot tables can be used to threats and keep them from causing
prepare data collection and create data further harm. The proposed system can
summaries. Excel also includes protect networks while also increasing
graphical representations of data, such security and removing security threats.
as pie charts, which can be used to
show the percentage of responses. Conclusion
Following the analysis of the data, final As stated in the research, the
conclusions and statements can be development of Artificial Intelligence
made and documented in docx. opens the door to a plethora of
opportunities that can be combined to
detect and prevent threats. Furthermore,
the findings conclude that information
technology is growing and evolving,
which causes problems such as an
Overview of the proposed system
increase in cyber-crime activities. In
summary, the study discusses the use
This research specializes in developing
of Artificial Intelligence strategies, their
an intelligent system using Artificial
scope for destiny work, and how they
Intelligence to hit upon/detect and
can be used to avoid large losses.
 [5] R. Trifonov, S. Manolov, R.
Yoshinov, G. Tsochev, and G.
Pavlova, "Artificial
Intelligence Methods for Cyber
Threats Intelligence," vol. 2,
2017, pp. 129–135.

[6] M. Brundage et al., "The

Reference Malicious Use of AI:
Forecasting, Prevention, and
Mitigation," no. February
[1]Application of artificial 2018, 2018.
intelligence for detecting
derived viruses," O. Asiru, in [7] J. Pan and C. Fung, "Artificial
European Conference on intelligence in malware-Cop or
Information Warfare and suspect?" December 2008, pp.
Security, ECCWS, 2017, no. 181–184.
July, p. 91.
[8] I. A. Saeed, A. Selamat, and A.
[2] S. Dilek, H. Cakr, and M. Aydn, M. A. Abuagoub, "A Survey
"Applications of Artificial on Malware and Malware
Intelligence Techniques to Detection Systems,"
Combating Cyber Crimes: A International Journal of
Review," International Journal Computer Applications, vol.
of Artificial Intelligence 67, no. 16, pp. 25–31, 2013.
Applications, vol. 6, no. 1, pp.
21–39, 2015. [9] E. Gandotra, D. Bansal, and S.
Sofat, "Malware Analysis and
[3] Z. Siddiqui, M. S. Husain, and S. Classification," April 2014, pp.
Yadav, "Application of 56–64.
Artificial Intelligence in
Fighting Cyber Crimes: A [10] D. Gavriluţ, M. Cimpoeşu, D.
Review," International Journal Anton, and L. Ciortuz,
of Advanced Research in "Malware detection using
Computer Science, vol. 9, no. machine learning," Proc.
2, pp. 118–122, 2018. International Multiconference
Comput.
[4] J. S. Mohan and T. Nilina,
"Prospects of Artificial
Intelligence in Tackling Cyber
Crimes," International Journal [11] G. L. Lowhorn, “Qualitative
of Scientific Research ISSN and Quantitative Research:
(Online Index Copernicus How to Choose the Best
Value Impact Factor, vol. 14, Design,” Pap. Present. Acad.
no. 6, pp. 2319–7064, 2013. Bus. World Int. Conf., no.
2003, pp. 1–5, 2007.
[12]. Bairagi, V. and Munot, M.V. eds.,
2019. Research methodology: A practical
and scientific approach. CRC Press.

[13]. Mishra, S.B. and Alok, S., 2017.

Handbook of research methodology.

[14]. Kumar, R., 2018. Research

methodology: A step-by-step guide for
beginners. Sage.

[15.] Ghadge, A., Weiß, M., Caldwell, N.D.

and Wilding, R., 2019. Managing cyber risk
in supply chains: A review and research
agenda. Supply Chain Management: An Appendices
International Journal.

Methodology Flowchart