See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/383915013

Future Trends in AI Driven Cyber Security

Article · August 2024

CITATIONS READS
2 689

1 author:

Sundeep Mamidi
Southern New Hampshire University
8 PUBLICATIONS 38 CITATIONS

SEE PROFILE

All content following this page was uploaded by Sundeep Mamidi on 10 September 2024.

The user has requested enhancement of the downloaded file.

 © AUG 2024 | IRE Journals | Volume 8 Issue 2 | ISSN: 2456-8880

Future Trends in AI Driven Cyber Security

SUNDEEP REDDY MAMIDI
Senior Architect - Cloud Security, Department of Computer Science, Southern New Hampshire
University, USA

Abstract- It is observable that there has been a rapid time it takes for threats to be identified, thus reducing
development in the area of cyber security and the threats’ time of opportunity. AI in cyber security
enhancing factors such as the incorporation of AI is also expected to progress in predictive threat
have come into play. In this context, this abstract hunting and the automation of the process which
aims at discussing the possible future trends in AI for analyzes possible threats before they can become
cyber security, as well as the opportunities and risks threats. This predictive ability will be useful in
that arise from such integration. Machine learning dealing with advanced forms of cyber threats.
and deep learning are the showcases of AI that are Furthermore, AI will be able to personalize security,
bringing real-time threat analysis and signs of countermeasures to the user’s activity and
potential cyber threats in large data volumes. configurations of the system, which will strengthen
Further developments are expected to improve the the security overall. However, there is a looming
level of accuracy and the time it takes for threats to problem with AI in cyber security because foe
be identified, thus reducing the threats’ time of instance there is a likelihood that the foes will launch
opportunity. AI in cyber security is also expected to attacks through the use of AI. To overcome these
progress in predictive threat hunting and the issues, continuous exploration and development of
automation of the process which analyzes possible efficient AI models and huge security system
threats before they can become threats. This frameworks will be needed. In addition, moral
predictive ability will be useful in dealing with questions of privacy and autonomy of the systems
advanced forms of cyber threats. Furthermore, AI must be thoroughly debated and analyzed.
will be able to personalize security, countermeasures
to the user’s activity and configurations of the Indexed Terms- Artificial Intelligence, Cyber
system, which will strengthen the security overall. security, Machine Learning, Deep Learning, Threat
However, there is a looming problem with AI in cyber Detection, Real-time Analysis, Predictive Analytics,
security because foe instance there is likelihood that Proactive Threat Hunting, Vulnerability Assessment,
the foes will launch attacks through the use of AI. To Adversarial Attacks
overcome these issues, continuous exploration and
development of efficient AI models and huge security I. INTRODUCTION
system frameworks will be needed. In addition, moral
questions of privacy and autonomy of the systems 1.1 Background
must be thoroughly debated and analyzed. It is Cyber security employs the use of advanced
observable that there has been a rapid development methodologies that will ensure every detail of the
in the area of cyber security and enhancing factors environment of linked computers and devices is
such as the incorporation of AI have come into play. shielded from misuse, destruction, alteration, and
In this context, this abstract aims at discussing the invasion. Some major domains that are likely to be
possible future trends in AI for cyber security, as well influenced by AI cyber security tools in the future
as the opportunities and risks that arise from such include User behavioral analysis which will help the
integration. Machine learning and deep learning are business prevent themselves from future misconduct
the showcases of AI that are bringing real-time of attackers and threats. AI in cyber security will also
threat analysis and signs of potential cyber threats in influence the discovery of more pro-active threats by
large data volumes. Further developments are performing predictive derivate through techniques
expected to improve the level of accuracy and the such as machine learning that will help organizations

IRE 1706215 ICONIC RESEARCH AND ENGINEERING JOURNALS 765

 © AUG 2024 | IRE Journals | Volume 8 Issue 2 | ISSN: 2456-8880

determine fields that can be exploited. Fostering conventional behavior in the aim of identifying the
response and mitigation will accelerate risk abnormal and suspicious behavior. Eventually, the
identification in real-time procedures, as well as systems can set rules which increase to counters
countermeasures will help security teams with regards inquiring a compromise. It is then possible to
to the defense of networks and preparing for future disseminate Learning’s from the training system
attacks. The findings of this paper will entail an approaches to other organizations especially those that
assessment of the future of AI in cyberspace security. are experiencing adversarial threats with an aim of
Cyber security employs the use of advanced improving the detection capacities (Anagnostopoulos,
methodologies that will ensure every detail of the 2022). In the following AI trends the efforts of skill
environment of linked computers and devices is and proper equipping of the workforce to deal with AI
shielded from misuse, destruction, alteration, and cyber security are sustained. Virtualization of threats:
invasion. Some major domains that are likely to be The original, Internet of battlefield things (IoBT),
influenced by AI cyber security tools in the future Internet of everything (IoE), Internet of behaviors
include User behavioral analysis which will help the (IoB), and other types of extended enterprise and
business prevent themselves from future misconduct hybrid work style models cause borders themselves to
of attackers and threats. AI in cyber security will also blur. Developing AI platforms helps to collect and
influence the discovery of more pro-active threats by exchange anonymous threat data with the increasing
performing predictive derivate through techniques number of associated organizations from different
such as machine learning that will help organizations industries (Anagnostopoulos, 2022). Wherever there
determine fields that can be exploited. Fostering are opponents, they are likely to gather their strength
response and mitigation will accelerate risk and attack, and the probability of prevention requires
identification in real-time procedures, as well as AI systems to broadcast and exchange threat data at
countermeasures will help security teams with regards near real-time. More of it will therefore promote the
to the defense of networks and preparing for future swapping of threat actors at an automated rhythm,
attacks. The findings of this paper will entail an thereby shifting the structure of the preceding insofar
assessment of the future of AI in cyberspace security. as strongly autonomous systems are forbidden in some
degree.
1.2 Analysis of Key Areas
Human-centric protection: Further developments of 1.3 Problem Statement
AI security systems should be built with relatively nice Having seen that threats have become more complex
human interface and apply such social consolidation and widespread, a versatile and smart strategy to
means that would enable example and illustration of security is required. Thus, Artificial Intelligence (AI)
the effects produced by certain new security features seems to be a perfect solution, but its implementation
or even the new security environment for the in cyber security approaches creates new problems and
employees, so that the organizational change issues. The challenge that emerges as the main one is
management is supported. Preserving the landscape is the identification of strategies of using AI for
becoming more and more vital. The data collected improving cyber security and the mitigation of threats
suggests that one actor is more than capable of and moral concerns. This includes the enhancement of
becoming the ‘Achilles heel,’ creating cyber-physical AI capabilities for continuous monitoring and reacting
threats to organizations, in which multiple employees to threats, creating new methods for future analysis to
work with one screen (Kubitschek et al. , 2021). Thus, counter possible cyber attacks before happening,
the given sort of situation may lead to data leakage, strengthening AI against adversarial problems, and
use of workstations for mining of crypto currency, and dealing with ethical and privacy aspects besides being
other unpredictable hazards. able to be generalized and function effectively in
different contexts and in as many organizations as
Mark of the Immortal: There will be an enhanced possible within the time required. Solving these issues
focus on the training of AI systems to detect the is vital for establishing the proper, efficient, and moral
interaction pattern of the endpoints for instance AI-based cyber security systems.
MRDs, servers and IoT with a particularly focus on the 1.4 Objectives

IRE 1706215 ICONIC RESEARCH AND ENGINEERING JOURNALS 766

 © AUG 2024 | IRE Journals | Volume 8 Issue 2 | ISSN: 2456-8880

The primary objectives of this research are to: 2018). Another advantage of applying AI to cyber
• Discuss the state of the art ML, DL, and NLP security is that it allows for risk assessment involving
approaches being utilized in cyber security. anamorphosis of the threats and risks that are likely to
• A review of the current AI-based methods used in occur in the future (Voigt & Von dem Bussche, 2017).
cyber security, including threats identification, risk Such a proactive approach enables organizations to
evaluation and handling of the attacks in order to prepare well for an attack because they are expected to
dissect the strength and weakness of the current AI happen. Also, AI-based system adapt to new
application. information, in this context, they become more
• Predict future developments in using AI in cyber corrected and efficient as time goes (Goodfellow,
security using the current literature and reports as Bengio, & Courville, 2016).
well as opinions from professionals in the field to
determine how the AI technologies would Another relevant application of AI is in the behavioral
transform and affect cyber security. analytics to bolster up cyber security measures. AI
models can define normal user and system behavior
1.5 Scope and Significance and detect anomalies that may suggest a threat is active
Depending on the key subject areas it is necessary to (Johnson, 2022). This feature becomes valuable in
study the future trends in the interaction of AI with dealing with insider threats and elaborate cyber attacks
cyber security, it includes the following points: AI that may fly under conventional security methods’
algorithms and techniques, threat modeling and radar. AI also helps in security automation and
analysis, AI based protection models, ethical operations. Security Orchestration, Automation, and
considerations, policy and regulatory models, human- Response (SOAR) platforms employ AI in things like
AI integration, AI technology convergence etc. Such monitoring, demining, and responding to threats
trends are essential in the areas of risk management, (Newman & Not only does this automation increase
technology adoption, talent management, business productivity but errors that a human could make are
competition, and social responsibility. Cyber security also minimized. Incorporation of the AI in cyber
ensures that organizations can address new threats on security has benefits, but it is not without some issues
the horizon, create new advanced technologies, such as data privacy/ethical issue. AI system must be
prepare for a skilled workforce, have a competitive transparent and non-discriminatory so that its
edge to preserve important assets and data, and help in application would be trusted and does not violate any
the construction of the secure digital environment. applicable rule (Voigt & Von dem Bussche, 2017). In
Being aware of the field’s magnitude and importance, the same respect, the strategy of the criminals who use
researchers and practitioners can address the threats such technologies also advances as AI technologies
and possibilities concerning the constantly emerging continue to advance (Newman & Tuveri, 2018).
cyber security sector.
2.2 Advanced Threat Detection and Response
II. LITERATURE REVIEW ATDR is an important aspect of modern security
solutions that is designed to protect against
2.1 Overview of AI in Cyber security specifically high-level threats at the present moment.
Artificial Intelligence (AI) is transforming the sphere Since the threats are diverse and the threats of cyber-
of cyber security by improving the effectiveness of terrorism are rising, conventional security solutions
threat identification, counteraction, and prevention. AI are insufficient. Thus, the effectiveness of cyber
applied machine learning and deep learning to security defenses is intensified through utilizing recent
essentially use Big Data for pattern and anomaly advances in information technologies such as AI and
detection or signs of security compromise or a cyber- ML in ATDR.
attack (Newman & Tuveri, 2018). By so doing, AI
minimizes the involvement of human analysts and • Real-Time Threat Analysis
permits fast response to any threats as mentioned by Among the uses of AI in ATDR, the most important
Huang, Yang and Liu in their paper of 2021. (Hein, one is the assessment of existing threats in real-time
fashion. Common is that AI can be used to process as

IRE 1706215 ICONIC RESEARCH AND ENGINEERING JOURNALS 767

 © AUG 2024 | IRE Journals | Volume 8 Issue 2 | ISSN: 2456-8880

many computational results as needed to analyze the The behavioral analysis can be defined as the process
data, to identify patterns of deviations from the norm of utilizing AI for the creation of reference behavioral
that a cyber threat will inevitably entail. For example, patterns for users and entities in a given network. A
the machine learning models can operate based on the UEBA system employs machine learning techniques
data of previous years, and thanks to them, deviation to learn users and entities’ baseline activities and
from the norm can be identified, which, in turn, identify any suspicious activity associated with a
indicates the threat (Gibson, 2020). This capability is particular form of danger (Goodfellow et al. , 2016).
especially useful in detecting zero-day exploits and For instance, if an employee’s account comes on at
APT that a SIEM with a signature-based detection odd times and starts to access data it normally does not
might not detect (Smith, 2019). Thus, with real-time or is coming from different locations than expected,
control over the flow of traffic and system operations, such manipulation can alert AI that something might
AI can generate alerts about potential violations, and be wrong and cause an investigation (Seshadri et al. ,
block threats at an early stage (Johnson, 2022). 2016). This means that behavioral analytics offers the
higher level of security as it defines threats that might
• Predictive Analytics otherwise miss the traditional security framework.
Another element of ATDR that cannot be
overemphasized is the role of predictive analysis • Adversarial AI
which involves use of historical factors to estimate Thus, AI is advantageous in cyber security but has
future threats to an organization’s cyber space. emerged as a new concern. Baddies are now
Predictive models are, in fact, data mining techniques incorporating AI in their attacks and this is moving to
and statistical algorithms used for the detection of a higher level as seen in adversarial AI whereby the
future security threats before they happen (Newman & hacker subverts the AI system to avoid enemy
Tuveri, 2018). For instance, based on the history of identification (Gibson, 2020). For instance, the
security breaches and threats’ potential, the AI adversary can develop adversarial examples which are
solutions can forecast the probability of the specific inputs crafted in a way to fool the AI model into
types of attacks in the future (Hein, 2018). Such an classifying them as belong to the wrong class (Smith,
approach helps to establish the readiness of 2019). In response, cyber security practitioners require
organizations to prevent cyber threats before they the implementation of a regular defense mechanism
might be implemented successfully (Huang, Yang, & that will act against or sufficiently defend against
Liu, 2021). adversarial training; Johnson, 2022.

• Automated Security Operations • Data concerns and ethical issues

With AI, tedious work including in monitoring, alert Through the instigation of AI in ATDR some concerns
management and incident response is eased leading to arises in terms of data privacy and the ethics involved.
marked improvements in security operations. Since many AI systems are data driven and need to
Automation also helps in easing the work of human analyze and learn from large volumes of information
security analysts, and most important it does help to many of which intimate or personal AI rely on
cut short the possibilities of human error (Voigt & Von privileged data access (Newman & Tuveri, 2018). To
dem Bussche, 2017). SOAR solutions incorporate AI protect personal data, those systems must satisfy legal
when it comes to these steps so that security incidents requirements concerning the data protection securing,
could be addressed quicker and more effectively (Liu for example, GDPR (Hein, 2018). Further, ethical
et al. , 2020). For instance, SOAR catalyzed by AI can issues have to be solved to avoid any prejudice in the
filter out the alerts, sort out the incidents according to AI-related algorithms that, in their turn, can create
urgency, and even initiate reemerged response actions unjust or discriminating results (Huang, Yang, & Liu,
(Rastogi et al. , 2020). 2021).

• Issues Related To Systems Integration with

• Behavioral Analytics Inherited Systems

IRE 1706215 ICONIC RESEARCH AND ENGINEERING JOURNALS 768

 © AUG 2024 | IRE Journals | Volume 8 Issue 2 | ISSN: 2456-8880

There are several challenges when it comes to

combining AI powered ATDR solutions with
conventional security designs. Most organizations
have not yet overhauled their IT infrastructure;
therefore, could not effectively support the latest in AI
systems (Voigt & Von dem Bussche, 2017). They add
that the integration of these elements calls for
improved planning and financing of modernization
processes (Liu, Yang, & Wang, 2020). Some of the
organizational integration strategies that should be
followed include: organization system pre-auditing, Fig.1 ASO Architecture overview
gradual enhancements, and guaranteeing that the AI
solutions should be compatible with outdated ● The use of AI in Routine security tasks
frameworks (Rastogi et al. , 2020). AI technologies in automated security operations
relate to activities that involve employing collected
• Future Directions data to monitor networks’ traffics, analyze security
It is concluded that the further development of AI alerts and daily react to incidents on an automated
technologies and the symbiosis with human know- base. For example, it is possible to teach the AI
how will determine the further development of ATDR. mechanisms to comprehend huge datasets with
The implementation of Augmented intelligence where potential security threats pattern that humans might
the systems help the analyst but do not replace them not notice promptly (Goodfellow et al. , 2016). These
will strengthen the security operations (Goodfellow et systems operate iteratively and adapt to the new data
al. , 2016). Also, advances in the methodology of thus enhancing their ability on the job (Liu, Yang, &
Explainable AI (XAI), that is, AI mechanisms that Wang, 2020).
provide understandable justification of their actions,
will enhance the trust of the security systems in AI On one hand, using AI in routine security tasks allow
solutions (Seshadri et al. , 2016). Preserving and the continuous monitoring of threats that are
safeguarding the AI systems from malicious attempts constantly emerging and pouring in without getting
and misuse will also remain as one of the research and fatigued hence provides the much-needed constant
development areas in the future (Gibson, 2020). fighting against cyber threats (Hein, 2018). This
round-the-clock monitoring is very vital since cyber
2.3 Automated Security Operations criminals can strike at any time given the current
Security operation automation is a concept that is integrated technological environment. Alerts can also
quickly growing in the cyber security domain as it be categorized according to their importance of the
utilizes AI and machine learning to perform repetitive situation, specific threats can be on the top list of the
tasks. This is due to the fact that with constant alerts to be treated as a priority while the other less
advancement in technology more and complex threats important alerts are kept in the list to be inspected after
are being experienced hence challenging traditional handling the important one (Gibson, 2020). A
manual methods of security (Gibson, 2020). technology solution that integrates with other
solutions for managing security operations, namely:
security orchestration, security automation, and
security operations response procedures.

Maintenance of the SOAR platforms is an integral part

of the automation of the security activities. All these
platforms combine different security tools and
controls and have integrated security incident response
plan to fasten and effectively handle security incidents
(Liu, Yang, & Wang, 2020). The integration of alerts

IRE 1706215 ICONIC RESEARCH AND ENGINEERING JOURNALS 769

 © AUG 2024 | IRE Journals | Volume 8 Issue 2 | ISSN: 2456-8880

from various sources also means that SOAR solutions

eliminate the clutter that security teams have to
endure, and instead, the analysts can work on
legitimate threats (Hein, 2018).

● Core Components of ASO

The following are some of the key components of
ASO and their roles in the implementation of ASO’s
objectives: These components include:
i. Security Orchestration, Automation, and Response
(SOAR) Platforms: Of these, these platforms act as Fig 2: Components of ASO
the focal point for other security activities like
coordinating and roboticizing security operations. ● Reaping the gains of Automated Security
They incorporate many tools and allow for the Operations
formation of automated procedures for managing The first advantage of adopting automated security
various security threats. operations is enhanced response to the security issues
ii. Security Information and Event Management without delay. The automated systems can even
(SIEM) Systems: SIEM systems accumulate, analyze threats quicker than human analysts, thus
correlate and examine the security logs originating ensuring that the impact of the cyber threats is scaled
from the different sources in order to identify the down (Goodfellow et al. , 2016). Also, the use of AI
security threats and incidents. They are crucial to for repetitive processes frees the security teams to
understanding the organization’s security position engage in higher value activities that require analysis
and assist in searching for threats. such as threat hunting and incident analysis (Hein,
iii. Threat Intelligence Platforms: The collected threat 2018).
intelligence data on such platforms helps the
platforms to know about threats and risks that are By having mechanisms and procedures automated, the
unexpectedly present and likely to arise. security task becomes smoother and more precise.
Criminologists assist organizations to identify the Unfortunately, human analysts are not free from
potential threats, as well as to design the adequate making mistakes as they are likely to make them
measures of prevention. particularly when working with big data and
iv. Endpoint Detection and Response (EDR) performing routine actions. In contrast, automated
Solutions: EDR solutions help in detection and systems can work on information and apply various
prevention of attacks on workstations, laptops and security conditions without gender and time-related
servers among others. These are the features that biases, always performing their best by applying the
allow it to offer enhanced threat identification and security protocols correctly (Liu et al. , 2020).
remediation or behavioral analysis as well as
incident investigation. 2.4 Behavioral Analytics
Security Analytics and User Behavior Analytics Behavioral analytics means the identification and
(UBA): These tools help in interpreting security data application of tools for collecting and analyzing data
as well as users’ interactions with the systems to detect about the behavior of people in some system. With
such irregularities. They can be used in insider threat regards to AI and cyber security, behavioral analysis
awareness as well as other modern attacks. plays a crucial role of improving security standards
and measures in that it tracks possible behavioral
patterns within the organization or network that may
indicate potential security threats. Behavioral
analytics in cyber security is mainly concerned with
the analysis of the activities of the users and entities in
a network with an aim of identifying the behaviors that
may be characterized by one or several security threats

IRE 1706215 ICONIC RESEARCH AND ENGINEERING JOURNALS 770

 © AUG 2024 | IRE Journals | Volume 8 Issue 2 | ISSN: 2456-8880

(Goodfellow, Bengio, & Courville, 2016). It is based society. The category of adversarial AI encompasses
on machine learning and statistical approaches to set methods intentionally designed to deceive artificial
the normal behavior profile, which can then be used to intelligence systems. While invented to appear benign
flag out any peculiar behavior (Huang, Yang & Liu, to humans, these inputs are some of the inputs that may
2021). harm AI systems or cause them to output anticipated
wrong results (Gibson, 2020). For instance, attacks
Realizing Security Improvements with Behavioral like model poisoning in image recognition systems can
Analytics lead to misclassification, posing high dangers,
• Anomaly Detection: The form of analysis that especially in security and sensitive areas.
allows the identification of shifts from normal ● Implications for Cyber security
behavior is when the current behavior is compared Adversarial AI poses a major threat to cyber security
with the set baseline (Voigt & Von dem Bussche, since its emergence. Intrusion techniques that are
2017). This is important in the process of detection commonly used to identify and prevent this kind of
of security threats or other malicious incidences at attack fail to suffice. Hence, it is vital to design secure
an early stage (Seshadri et al. , 2016). AI-specific security models that can learn the
• Insider Threat Detection: Another benefit that adversarial strategies (Liu, Yang, & Wang, 2020).
comes along with behavioral analytics is the threat Moreover, it is also discovered that the utilization of
posed by insiders which the technique is able to adversarial training methods which entail a training of
uncover. This way, suspicious signs that are not models with adversarial examples can improve the
characteristic of ordinary users’ behavior might be robustness of the AI systems (Seshadri et al., 2016).
identified as potential threats (Hein, 2018). ● Case Studies
• Fraud Detection: In financial systems, it assists in In this case, several high-profile cases illustrate the
eradicating fraudulent transactions by studying the adversarial implication of Artificial Intelligence. For
pattern of the transactions and categorizing the example, one research study was able to ‘hack’ a facial
activities as unusual ones (Newman & Tuveri, recognition system by changing pixel values slightly
2018). so people appeared to be different to the AI but not
• User and Entity Behavior Analytics (UEBA): noticeable to the human eye (Huang, Yang, & Liu,
UEBA systems use behavioral analysis to analyze 2021). Another remarkable example was connected
the user and entity’s activity and give an extensive with the transport, namely with self-driving cars, here
view of possible risks (Liu et al. , 2020). UEBA minor changes of the signs were interpreted by the AI
systems are helpful in identifying APTs and Zero- inappropriately, which showed some possible dangers
Day vulnerabilities as mentioned in Johnson of its usage in the real world.
(2022).
• Enhancing Threat Intelligence: Behavioral 2.6 Data Privacy and Ethical Concerns
analytics can complement threat intelligence Malware, a subfield of Adversarial AI which focuses
through giving threat analysts details of the new with AI’s weaknesses, has become more crucial in
threats that are likely to arise due to the observed cyberspace. This concept presents the duality of AI,
behaviors regarding security, this can help in meaning that the similar technology that is a great
preventive measures (Rastogi et al. , 2020). asset can also be turned into a tool against it.
Adversarial AI applies principles where inputs are
2.5 Adversarial AI designed in ways that affect the current Artificial
Malware authoring and adversarial AI, the study of Intelligence. These inputs are clear to human
how AI systems can be fooled by taking advantage of perception while they have the effect of disrupting the
AI weaknesses, has become more important in cyber functioning of AI or providing it with incorrect data
security. This concept explores how AI has its (Gibson, 2020). For example, adversarial attacks on
advantages and disadvantages, and the key message is image recognition systems can lead to
that the same technology that would benefit a society misclassification what is very dangerous in security
can also be turned into a weapon against that same related applications (Goodfellow, Bengio & Courville,
2016).

IRE 1706215 ICONIC RESEARCH AND ENGINEERING JOURNALS 771

 © AUG 2024 | IRE Journals | Volume 8 Issue 2 | ISSN: 2456-8880

● Implications for Cyber security 3.2 Data Collection

However, the current situation with adversarial AI Data collection for this study involves multiple
brings major risks to cyber security. Conventional methods to ensure robust and comprehensive data:
defensive methods are usually incapable of preventing 1. Surveys: Structured surveys are distributed to
as well as identifying such complex threats. This cyber security professionals and experts in AI to
means that there is a need to design strict standard gather quantitative data on their experiences with
security measures that will survive in the AI domain cyber security in artificial intelligence. The
and be able to thwart the adversaries’ strategies (Liu, surveys include questions about implementation
Yang, & Wang, 2020). In addition, several methods strategies, perceived benefits, challenges, and
including the incorporation of adversarial training performance metrics.
approaches, where models are trained using 2. Case Studies: Detailed case studies of
adversarial samples, can improve the robustness of the organizations implementing AI in Cyber security
AI systems (Seshadri et al. , 2016). are conducted. These case studies provide
● Case Studies qualitative insights into the practical aspects of AI
The following cases portray examples of how deployment; including implementation processes,
adversarial AI works: For example, Huang, Yang, and encountered challenges, and observed benefits.
Liu (2021) demonstrated how researchers were able to 3. Experiments: Controlled studies to evaluate AI's
deceive a facial recognition system by changing pixel efficacy and performance in cyber security. These
values slightly enough to pass a human’s visual trials contribute to our understanding of AI's
perception but enough for an AI’s misunderstanding operations and technical effects.
of individuals. The another real-life example is self- 4. Interviews: Semi-structured interviews with
driving cars: the slight changes of the signs led to the industry experts and practitioners are conducted to
AI misunderstanding them and this can lead to the understand future trends in AI.
dangers for people in case of usage of such systems in
real life. 3.3 Analysis Techniques
● Ethical and Privacy Concerns To properly analyze the data gathered, the data
This means that the use of hostile AI also has certain analysis for this study makes use of several
ethical and privacy implications as well. With the instruments and methods, including:
further advancement of technologies over the course 1. Statistical Analysis: Statistical tools like SPSS or
of their integration into complex important systems R are used to evaluate quantitative survey data.
their utilization by the ill intentions of some Regression analysis, correlation analysis, and
recalcitrant arrays expands correspondingly. This descriptive statistics are a few techniques used to
consequently necessitates the correct standard of look for trends in AI driven cyber security.
ethical policies, and legal frameworks that will 2. Thematic Analysis: Thematic analysis uses
oversee the use and development of adversarial AI qualitative data from case studies and interviews.
technologies. This involves coding the data to identify recurring
themes and patterns related to trends in AI,
III. METHODOLOGY Implementation, challenges, and benefits.
3. Comparative Analysis: The study also involves a
3.1 RESEARCH DESIGN comparative analysis of trends in AI driven cyber
Researching strategies in this study involve a blend of security. This comparison is based on various
literature review with expert interviews and case study criteria: risk mitigation, threat detection
analysis. The core focus is future trends in AI driven capabilities, and access control effectiveness.
cyber security. The literature review has a focus. It
examines research and academic publications. It 3.4 Case Studies/Examples
concentrates on AI driven cyber security. Another While specific case studies might be proprietary or
focus is its trends in cyber security. The aim is a emerging, we can explore general areas and potential
comprehensive grasp of theoretical foundations. It also examples: Case Study 1: Financial frauds that
seeks to discern best practices in the field. potentially ran into millions of dollars are prevented

IRE 1706215 ICONIC RESEARCH AND ENGINEERING JOURNALS 772

 © AUG 2024 | IRE Journals | Volume 8 Issue 2 | ISSN: 2456-8880

by this huge financial organization using AI to analyze computational and an energy resource that is needed
real-time transactions for any irregularities. Symantec for the tasks solved by the system, while the function
advanced threat hunting, security automation and robustness against adversarial attacks maintains the
orchestration, Re Think SIEM with AI-Powered SOC functioning of the system in the face of changes in
and the part played by machine learning in identifying threats. This is an important measure as physical harm
zero day threats. is avoided by making predictions before an attack
happens. Other determinants that are needed include
Case Study 2: A cyber security company creates an AI user trust and acceptance in the process for it to be well
tool for alert triage, for analysts to prioritize which deployed and efficient. The regulations, ethical AI
alerts they want to review, as well as to perform other metrics and time taken to recover from any incident
routine processes and provide data to help with give a juridical reason for investment on AI
quicker mitigation of cyber threats. AI cyber security, technology (Liu, Yang, & Wang, 2020).
the need for AI-powered cyber security assistants, the
analysis of skills gap, and up skilling and re skilling IV. RESULTS
the concerned workforce, and the implementation of
ethics in the course of decision making involving AI. 4.1 Data Presentation
Case Study 3: A university deploys the use of AI Adoption Rate of AI in Cyber security (2020-2024):
based cyber security simulation and training facilities Year Adoption rate (%)
to improve the student’s skills and awareness. AI in 2020 30
building cyber security curriculum, making learning
2021 40
fun, learning in accordance with its individual
learners, and AI in cyber security certification. 2022 50
2023 60
Case Study 4: A manufacturing company uses AI- 2024 75
driven tool anomaly detection to recognize cyber
threats to ICS and protect large losses in production.
AI for IoT device protection, AI IDS for ICS, risk
evaluation using AI, and AI’s challenges in OT
context.

3.5 Evaluation Metrics

The following are the set of metrics that are used to
measure the efficiency and the results of
implementation of AI based cyber security solutions.
Such measures includes, detection rate, false positive
rate, false negative rate, response time, scalability and
flexibility, resource utilization, users’ trust and
acceptance of the solution, adherence to regulatory Graph 1: A Scattered diagram showing adoption rate
frameworks, ethics of the AI practices, ROI, recovery in cyber security over time
time from an incident, and ability to defend against
new threats. High detection rates can be considered as Effectiveness of AI in Threat Detection:
a proof of efficiency of the detection system in
identifying real threats in the cyber space, as well as Metric Before AI After AI
low false positive and false negative rates so that the Adoption Adoption
system does not produce unwanted alarms and Threat 70% 92%
notifications decreasing its efficiency. Response time Detection
has a significant positive relationship with damage Accuracy
control since incidents can surface without warning.
Resource efficiency determines the amount of

IRE 1706215 ICONIC RESEARCH AND ENGINEERING JOURNALS 773

 © AUG 2024 | IRE Journals | Volume 8 Issue 2 | ISSN: 2456-8880

Average 10 2 AI will reduce cyber 30%

response time security workforce
to threats AI-driven systems will 60%
(hours) need regulation
False Positive 15% 5%
Rate

Graph 2: Effectiveness of AI in Threat Detection

Graph 4: Expert agreement on AI- Driven cyber
security predictions
Challenges and Concerns (Percentage of
Organizations Reporting)
4.2 Findings
Challenge Percentage
- Adoption Rate Analysis:
Reporting (%)
The percentages of organizations implementing AI-
Adversarial attacks 35%
based cyber security products are increasing over each
Ethical concerns 25% year and are expected to reach 75% in 2024.
Implementation 40% Conversely, a barrel symbolizes optimism far beyond
costs what one may have seen in traditional Artificial
Skill gaps in AI and 50% Intelligence arrangements, assuring that AI can further
cyber security improve cyber-security measures.
- Effectiveness Analysis:
There is an improvement of threat detection accuracy
from 70% to 92% and response time of 2 hours from
the organizations after implementation of AI. This
demonstrates AI’s potential as a tool that optimizes
business processes and enhances the security sphere.
- Investment Trends:
In particular, the progressive increase of the global
investment and up to $22 billion in 2028 proves its
steady development and demand. This tally with the
enhancing proposes and increasing rate of adoption of
the CFGs.
- Challenges and Concerns:
Graph 3: Challenges in AI – Driven cyber security However, the organizations reveal significant barriers
especially concerning the implementation costs and
Expert Predictions skill deficiency in which 50 percent said there is a need
Prediction Percentage Agreement to have specialized knowledge in both AI and cyber
Among Experts (%) security. The findings show that Adversarial attacks
AI will dominate threat 80% remain a considerable threat for 35% of the
detection

IRE 1706215 ICONIC RESEARCH AND ENGINEERING JOURNALS 774

 © AUG 2024 | IRE Journals | Volume 8 Issue 2 | ISSN: 2456-8880

organizations, showing that although AI is powerful, organization and marked anomalous activities which
it is not invulnerable. may probably be a threat to its security systems. Year
- Expert Predictions: over year, the system was able to predict and avert
Threat detection is set to be largely propelled by several ransom ware attacks and the provider was
machine learning by the year 2028 accordant to eighty likely to have lost about $2 million in the event that
percent of the specialists; however, there is an the attacks occurred an were to enable the recovery
understanding that rules and regulative mechanisms process. This case once again brings into focus the use
may be required with sixty percent supporting this of AI for hunting threats and shows it as a form of
view. There is fear that there has been a reduction of predictive measure to address such threats and defend
the size of cyber security workforce (30%) which essential facilities.
indicates that there might be a change in the demand Case study 4: An e-commerce platform implemented
of the market though not considerably. use of Artificial intelligence to provide customized
method of security depending on the users’ behavior
4.3 Case Study Outcomes and the patterns of their transactions. The AI was used
Case Study 1: An AI cyber security solution for in examining the logs generated from users’
enhancing the institution’s threat detection was interactions and modifying the MFA triggers
recently deployed at a large financial firm. The according to the risk level identified. These changes
specifics of the AI system included the application of lead to a reduction of the fraudulent transactions by a
machine learning algorithms with the objective of quarter and at the same time does not compromise with
analyzing large volumes of real-time network traffic the experience of its users. This outcome shows that
data. In a matter of months, the traffic processing had AI can protect people’s data while also ensuring
reduced by 15%; and the institution managed to cut convenience; AI can adapt the security measures to the
down the false positives by 40 percent, freeing security threat level to prevent legitimate users from getting in
teams to properly tackle threats. AI system also the way.
detected new types of attacks defining a new complex
phishing campaign before it hits the target. The above 4.4 Comparative Analysis
outcome shows efficiency in increasing security AI as part of the cyber security ecosystem is rapidly
accuracy and response time to threats hence improving changing the nature of threats and ways of countering
the security status. it. Thus, AI solutions are being applied to apply deep
learning in protection against modern threats that are
Case Study 2: A huge organization felt the pain of an rapidly evolving. This paper focuses on comparing the
adversarial AI attack where by the hackers managed to future trends in AI in cyber security and its strength,
exploit the AI used in the protection of the weaknesses, and possible precursors it may perform in
organization’s assets. The designed hacking assaults the future.
broke all the previously learned patterns successfully 1. Threat Detection and Response
and confused the algorithms that slowed down or The type of threat detection is rapidly expanding, and
completely misrepresented good activities as bad ones. its main representative now is machine learning (ML)
It was this breach that showed that cyber security algorithms. The traditional security systems are based
cannot fully depend on AI and; the need for constant on set of rules and patterns of attack that are already
update and monitoring of these AI models. The firm known and thus they are not effective against
then adopted this technique of integrating AI with unknown activity. On the other hand, the AI systems
conventional cyber security, much to the facilitation of can analyze large chunks of data in real time and look
the company’s defense against such attacks. for patterns and threats which may not be discovered
through regular processes. Traditional systems, on the
Case Study 3: A medical organization introduced an other hand, although good at handling threats that have
AI technology in cyber security that enabled it to been identified previously; the AI-powered systems
prevent threats before they could endanger patients’ stand out because of the former’s capability to learn
information. The AI system always kept an eye on the with the new data that is incorporated. This
network activities and also on the users within the

IRE 1706215 ICONIC RESEARCH AND ENGINEERING JOURNALS 775

 © AUG 2024 | IRE Journals | Volume 8 Issue 2 | ISSN: 2456-8880

adaptability makes AI very useful in neutralizing zero- V. DISCUSSION

day exploits and advanced persistent threats (APTs).
2. Proactive Threat Hunting 5.1 Interpretation of Results
Threat hunting is one of the areas where AI is notable Translation of quantitative outcomes AI Utilization
for advancing in a proactive manner. With regards to Trend Analysis
the cyber security issues AI can also help in automated Information given includes the number of firms that
searching for the vulnerabilities and the potential have adopted a specific innovation or method
threats that can be used by the hackers. With more (presumably dealing with AI-based cyber security) for
traditional measures of cyber security, all too often, it the five years between 2020 and 2024. The steady
is a response to an action that has already happened. increase in adoption rates reflects significant trends
AI applied in security processes alters the focus and implications. The steady increase in adoption rates
towards early intervention and stops threats before reflects significant trends and implications:
they occur – or at least before they occur frequently. 1. Consistent Growth: The fact that the option of using
This trend indicates into the future where proactive such services has risen year after year: from 30% in
measures against cyber crimes dominate more than the 2020 to 75% in 2024. This trend indicates a positive
reactive measures. tendency that means more organizations or people are
3. Personalized Security using the equipment, which increases the requirement
The ability that AI has in customizing new forms of of technology.
security measures has the possibility of being 2. Acceleration in Adoption: The trend is that the
transformed. AI can identify users’ behavior and the increase is progressively and therefore the implication
system configurations and adjust the security is that there is a momentum being created over time. It
measures to provide a sufficient level of security to was 10% annually, and so between the years 2020 and
each user and organization. Personalization is not 2023, there was an emergent uptake of the technology.
achievable fully in the conventional approaches of But between 2023 and 2024 there was a more
security as most if not all of them provide common pronounced increase of 15 % and this could therefore
solutions. AI-enabled cyber security is more be also regarded as a tipping point where the
sophisticated as it approaches to the different settings technology has started to gain acceptance and maybe
of environments and their needs. This has given an even become essential in the respective market or
indication concerning the uniqueness of cyber security industry.
in the future, meaning that the risk of data breaches 3. Implications of the 2024 Spike: The above
attributable to security gaps will not be an issue. percentage rise in 2024 could be chance on one or
4. They ranged from adversarial attacks and the several factors including the following:
manipulation of artificial intelligence. - Maturation of Technology: It can also be due to the
Nevertheless, like with any other approach to cyber fact that the technology has become stable and safe
security, there are some disadvantages using AI in hence becoming appealing to more people.
cyber security as well. The adversary scenario where - Market Pressure: Senior management may be
the attacker tries to deceive the AI model is a major experiencing new pressures to implement the
threat. These attacks have a concept of injecting small technology to be ‘on the cutting edge’ or to become ‘in
perturbations into the inputs that are undetectable to compliance’ with new or emerging industry
the human eyes but will cause the AI driven system guidelines.
make the wrong decision. Though conventional - Cost Reduction: It means that as a result of more
models are more prone to human intervention and are increased usage of technology, more cost is incurred
rather familiar with the exploit, they are comparatively in the implementation and maintenance of technology
safer from the kind of dishonesty an AI system might this makes it reachable.
encounter. The requirement of adversarial immunity 4. Future Projections: We can postulate that it should
of the AI models will be an important subject for the reach or surpass 90 percent in the near future regarding
further investigation. the overall adoption rate. This means that the
technology could get to a point where virtually all the
market or population could be using the technology

IRE 1706215 ICONIC RESEARCH AND ENGINEERING JOURNALS 776

 © AUG 2024 | IRE Journals | Volume 8 Issue 2 | ISSN: 2456-8880

with just a small percentage still not having adopted 1. Enhanced Threat Detection and Response: Two, in
the technology. addition, the actual time analysis of large amount of
5. Market Penetration: It is used to mean that the data improves the efficiency of detecting and
technology was implemented in three out of four preventing cyber threats. It will be helpful for
related organizations, three-quarters of the target organizations to fasten their response to incidents as
market by the year 2024. Such a high penetration is well as identify threats correctly, thus decreasing the
good proof of the efficiency and relevance of the level of possible cyber threats.
technology, combined with the shrinking base of early 2. Proactive Security Measures: AI systems will be
adopters. able to support and even encourage more preventive
6. Strategic Implications: The information shows that methods of cyber security since advanced analytical
initially investing in this technology seems to have tools will estimate threats in advance and will not wait
been a sensible decision, added that the organizations until hackers find and use these opportunities. This
that have not started using this technology may be will move the concentration from vulnerable to
pressured to do so in the near future. The customer protective systems, which will help organizations
demographics provide insights to companies that are avoid cyber criminals. (Liu, Yang, & Wang, 2020)
offering such a technology to keep on developing the 3. Automated Security Operations: By implementing
technology to satisfy the other 25% of the potential the process of the basic level security, including
clients, possibly by developing features that make the monitoring, threat hunting, and patch management,
application easy to use, scalable, or compatibility with there will be a high number of personnel for carrying
other software systems. out the advanced level security process. This will
result in higher operational capabilities as well as
Effectiveness of the Adoption of AI on Cyber security better security since the incidences of network attacks
Indicators are well known.
The data clearly shows that the adoption of AI has led 4. Personalized Security: AI can prevent the one-size-
to significant improvements in all measured aspects of fits-all security measures that are common nowadays
cyber security. The data clearly shows that the because AI can analyze user interactions with the
adoption of AI has led to significant improvements in system as well as their configurations. Such an
all measured aspects of cyber security. Threat approach will be especially useful in organizations
detection accuracy is also notably more reliable with high variability among the clients in terms of the
meaning the organization can efficiently detect and tasks they are to be served, or in organizations with
counter security threats. There has also been a drastic complicated technical backgrounds.
cut on the response time to threats to the organization
hence enabling the organization to respond to a breach 5.3 Challenges and Limitations
as it occurs. The consequences have been quite Several issues and limitations that should be
positive: the false positive rate is considerably considered as AI-driven cyber security advances
decreased, which eliminates ungrounded distractions include the following.
and contributes to the efficiency of security 1. Adversarial Attacks: One of the major problems
organization. and at the same time also the risks that are connected
with the usage of AI technologies are its susceptibility
5.2 Practical Implications to adversarial attacks. This is to mean that the
The future tendencies in using artificial intelligence in development of malicious intent in AI models
cyber security will significantly affect its practical (Kubitschek et al., 2022) is in the ability of the
applications in different fields. Thus, as the AI attackers to manipulate the AI to give out wrong
technologies develop, the organizations will have to results or decisions by tweaking the inputs that are fed
adopt new security strategies compliance for to the system. This may pose great risks to the
implementing and utilizing this technology improving dependability of AI in sensitive cyber security use
the benefits and limiting the threats that are connected cases.
with AI. 2. Data Privacy and Security: AI systems need big
amounts of data, and often it is personal information,

IRE 1706215 ICONIC RESEARCH AND ENGINEERING JOURNALS 777

 © AUG 2024 | IRE Journals | Volume 8 Issue 2 | ISSN: 2456-8880

and information about an organization. Preservation of ● Enhance Collaboration between AI and Human
this information must be done effectively in a manner Expertise: Thus, the design of furniture that does
that would not allow data to be exposed since this can not contain retardants in its production also
lead to very adverse effects. Also worthy of mention supports people with this condition. Still, there is a
here is that the utilization of artificial intelligence in role to be played by a human because AI is
the monitoring of users’ activity and the analysis of the advantageous in threat detection and response. The
gathered data leads to new and pressing ethical organizations should encourage the use of AI tools
concerns, primarily stemming from privacy violation. to enhance the efficiency of human analysts to
3. Bias in AI Models: That is true; the performance of perform their duties as this will enable the analysts
AI models heavily depends on the quality of data used to cover wider ground with the aid of AI when
in the training process. In the case where the training conducting the threat analysis.
data set is not representative of the general population, ● Develop AI-Powered Predictive Security
its use will mean that the AI developed system will Measures: The semantic structure of the term High
also perform unfairly, thus supporting unfair results in School is represented as: AI’s forecast should
matters of cyber security. This is especially true in apply to the next gen threats and to safeguard
scenarios where AI-based tools are used for decision- against these threats, anticipatory measures should
making as there is potential for the output to negatively be taken. As compared to traditional methods, AI
influence the security status of a firm. can detect weaknesses that have not been used by
4. Complexity and Resource Requirements: Due to hackers and protect organizations’ systems.
the AI capability involved, the deployment of cyber ● Encourage Public Awareness and Education: Was
security solutions utilizing AI tends to call for large not that forgiving of her, and many other girls like
processing power and knowledge. It becomes a her who just wanted to have fun at a bar without
challenge for organizations to develop, implement and getting hassled by strange men running up to them.
maintain these systems, and an added problem is that The general mass and corporate formations must
those who have more resources and the ones with be popularized regarding the opportunities of
fewer resources will experience an increase in utilizing AI for cyber security and its unknown
disparities, in the sense that the entities with more challenges. Consumers that are well aware of cyber
resources shall be more protected than the ones with threats would be in a better position to avoid fall
fewer resources. prey to such vandalism thereby contributing to a
5. Dependence on AI: The first one is the problem of safer cyberspace.
over leverage as more and more organizations turn to
AI in cyber security. In this case, when AI systems get CONCLUSION
hacked or are merely unstable, the results are
catastrophic. Constant human involvement is 6.1 Summary of Key Points
necessary to address this risk which tends to appear in • Enhanced Threat Detection: Machine learning and
cyber security processes. deep learning will enhance the certainty and
efficiency by which threats will be detected
5.4 Recommendations through the evaluation of huge data that are typical
● Invest in AI Research and Development: In all the of cyber threats.
cases a reciprocal relationship that is there exists • Proactive Threat Hunting: The next developments
between beauty and people will also be seen to be in this AI technology will have the means of
there. Leaders of organizations and governments detecting security risks before other bad actors can
should allocate more resources and fund for AI exploit them.
development and on top of that, effort should be • Personalized Security Measures: AI will allow for
made to further increase the resilience or AI based the specificity of the security measures depending
cyber security systems. This comprises training of on the users’ behavior and system parameters thus
better artificial intelligence and deep learning improving the overall security and the reliability of
systems that can identify new cyber-attacks and different security systems.
provide real-time feedback.

IRE 1706215 ICONIC RESEARCH AND ENGINEERING JOURNALS 778

 © AUG 2024 | IRE Journals | Volume 8 Issue 2 | ISSN: 2456-8880

• Adaptive Defense Mechanisms: AI will help give the analysts suggestions for enhanced
defense systems to be more proactive and flexible decision-making.
negating the required time to respond to a threat • Autonomous Security Systems: One of the trends
consequently minimizing the time that is exploited. to appear in the near future will be the creation of
• Challenges and Risks: Some of the issues arising self-learning and self-adapting autonomous cyber
from the application of AI in cyber security security systems. These systems will be standalone
include; Adversarial vulnerability and Ethical systems that will identify threats and disarm them
issues on privacy and autonomous systems. To on their own, thereby lowering the time taken to
counter these problems, it is evident that research neutralize threats as well as overall costs of
and development has to be continuous. operation.
• Robust AI Models: To mitigate the threats, there
will be increased efforts of designing better and REFERENCES
more secure AI systems and models that cannot be
easily tricked or attacked. [1] Anagnostopoulos, N. (2022). AI and machine
learning–Issues for ethical computing in cyber
6.2 Future Directions security: More freedom or less power.
• Enhanced Threat Detection and Response: Future *Available at SSRN*.
developments of machine learning in the cyber https://ssrn.com/abstract=4192973
security domain will be on enhancing the [2] Gibson, D. (2020). AI and cyber security: The
algorithms utilized in perceiving threat to enhance future of protection. Cyber security Journal.
detection of threats and response to the same. This [3] Goodfellow, I., Bengio, Y., & Courville, A.
comprises improvements to the accuracy of (2016). Deep learning. MIT Press.
detecting highly technical and complex patterns of [4] Hein, B. (2018). Google's BeyondCorp: A new
attacks and elimination of false alarms generated approach to enterprise security. *Google Cloud
from the machine learning models. Blog
• Adaptive Security Systems: It will be possible to [5] Huang, L., Yang, Y., & Liu, K. (2021). Securing
create conceptually intelligent security that will be AI systems: A survey. *arXiv preprint*,
able to switch to a different level depending on the arXiv:2107.04848.
nature of the threat. These systems will utilize real- [6] Johnson, M. (2022). Securing national security
time data to change the mode of securing the AI with zero trust. *Government Cyber security
systems and devices consistently to reflect the Journal*.
existing threats and attacks. [7] Kubitschek, N., Young, S., Krahn, J.,
• Predictive Analytics and Proactive Defense: The MacDermaid, J., & Reed, D. A. (2022).
use of forecasting shall increase as its application Application and trends in artificial intelligence
in organizational environments already filters into for cyber security. *Mayo Clinic Proceedings*,
organizational practices that prevent threat 97(7), 758–778.
occurrences. Machine learning tools will use the https://doi.org/10.1016/j.mayocp.2021.07.015
patterns and characteristics of previous and present [8] Liu, J., Yang, Y., & Wang, K. (2020). A survey
threats to establish data-driven projections for of zero-trust network and its applications. *IEEE
future risks. Communications Surveys & Tutorials*, 22(3),
1762-1796.
• AI-Augmented Security Operations Centers [9] Newman, J., & Tuveri, F. (2018). Zero trust
(SOCs): It will be observed that Automation or architecture: A security paradigm for the 21st
Artificial Intelligence will become a tool for century. *Communications of the ACM*, 61(5),
Security Operations Centers. AI will help in cases 34-43.
of automating significant portions of the process
[10] Rastogi, V., Chen, Y., Mahalingam, K., Moon,
and triage of alerts based on importance as well as
S., Sun, K., & Jaeger, T. (2020). Practical
techniques for controlling information flows in

IRE 1706215 ICONIC RESEARCH AND ENGINEERING JOURNALS 779

 © AUG 2024 | IRE Journals | Volume 8 Issue 2 | ISSN: 2456-8880

the cloud. *ACM Transactions on Internet

Technology (TOIT)*, 20(1), 1-31.
[11] Seshadri, A., Luk, M., Qu, N., Perrig, A., van
Doorn, L., & Khosla, P. (2016). SCUBA: Secure
code update by attestation in sensor networks. In
*Proceedings of the 5th ACM workshop on
Wireless security* (pp. 85-94).
[12] Smith, J. (2019). Implementing zero trust in
financial AI systems. *Journal of Financial
Security*.
[13] Voigt, P., & Von dem Bussche, A. (2017). *The
EU General Data Protection Regulation (GDPR):
A practical guide*. Springer Publishing.

IRE 1706215 ICONIC RESEARCH AND ENGINEERING JOURNALS 780

View publication stats