1

ARTIFICIAL INTELLIGENCE, DATA PROTECTION AND CYBER

LAW

RESEARCH PAPER

ON

“CONVERGENCE BETWEEN ARTIFICIAL INTELLIGENCE AND

ETHICS IN CYBERSECURITY”

SUBMITTED TO – Professor Amrit Subhadarsi

Prepared by- Prerna Bhansali

(20200401016)
Semester- 7-B
 2

Table of Content

Table of Content.....................................................................................................................................2
Abstract...................................................................................................................................................3
Introduction............................................................................................................................................4
Clarifying Key Concepts: Cybersecurity and AI................................................................................5
Deciphering Cybersecurity: Safeguarding Information and Assets.............................................5
Uncomplicating Artificial Intelligence: A Basic Insight.................................................................6
Empowering Cybersecurity Through Artificial Intelligence.............................................................6
Exploring the AI Agent Paradigm: Collaborative Intelligence in Action....................................7
Brain-Inspired Learning of Artificial Neural Networks (ANNs)..................................................7
Immune-Based Techniques in Cybersecurity.................................................................................8
Expert Systems in AI Applications...................................................................................................8
Leveraging AI Advantages in Cybersecurity......................................................................................9
AI Drawbacks in Cybersecurity.........................................................................................................10
New Frontiers of Cyber-Attacks on AI Systems...............................................................................11
The Ethical Quandaries of AI in Cybersecurity...............................................................................11
Conclusion............................................................................................................................................12
 3

CONVERGENCE BETWEEN ARTIFICIAL INTELLIGENCE

AND ETHICS IN CYBERSECURITY

Abstract
In the face of escalating cyber threats, cybersecurity has become a paramount societal
challenge, necessitating innovative solutions beyond traditional approaches. This paper
delves into the convergence of artificial intelligence (AI) and cybersecurity as a
transformative defense mechanism against evolving cyber-attacks. As conventional
cybersecurity methods prove inadequate, the integration of AI brings forth a powerful and
adaptive safeguard. However, this symbiotic relationship introduces ethical challenges, as the
very technology designed to fortify against cybercrimes may be manipulated by malicious
actors. The paper explores recent AI techniques implemented in cybersecurity while shedding
light on the ethical dilemmas intrinsic to this amalgamation. By dissecting the ethical
challenges associated with the intersection of AI and cybersecurity, the research aims to
contribute to a comprehensive understanding of the intricate dynamics and implications of
this critical paradigm shift in the realm of digital security.

Key Words: cybersecurity, AI, Machine Learning, ethics.

 4

Introduction
In recent times, there has been a growing focus on the development and application of
artificial intelligence (AI) and machine learning (ML). These technologies are being actively
explored across various industries,1 showcasing the unprecedented growth of AI research
driven by the global quest for technological solutions to society's most urgent issues.2

One such concern, highlighted in the 2019 Global Risks Report by the World Economic
Forum, positions cyber-attacks among the top five global risks with severe consequences.
This aligns with multiple analyses indicating a surge in the frequency and impact of cyber
threats. For instance, the first half of 2018 witnessed cyber-attacks compromising 3.3 billion
records, a significant increase of nearly 70% compared to the entirety of 2017 (2.7 billion).
These attacks are not only intensifying but also evolving rapidly. A Microsoft study reveals
that 60% of 2018's cyber-attacks lasted less than an hour, employing innovative forms of
malware.3

The financial toll of cybercrime is also on the rise, projected to double from US$3 trillion in
2015 to US$6 trillion in 2021 and reaching an estimated US$10.5 trillion by 2025. The
average cost of a single data breach soared to US$4.24 million in 2021, marking a 10%
increase from 2019. Ransomware claims alone have surged by 150% since 2018, according
to insurer AIG.

In response to this global challenge, both public and private entities, including corporations
and governments, are actively pursuing measures to safeguard their data and information.
Strategies include pseudonymization of data, data backup, and data encryption. The
integration of AI in cybersecurity aims to enhance prevention, identification, and resolution
of cyber-attacks, along with identifying vulnerabilities with minimal human intervention.
While AI adoption in cybersecurity is still in its early stages, the global market is anticipated
1
National Academies of Sciences, Engineering and Medicine, Implications of Artificial Intelligence for
Cybersecurity: Proceedings of a Workshop (The National Academies Press, Washington, DC 2019). Retrieved
from https://nap.nationalacademies.org/catalog/25488/implications-of-artificial-intelligence-for-cybersecurity-
proceedings-of-a-workshop. Published in 2019.
2
Brandon W Jackson. Minnesota Journal of Law. ‘Cybersecurity, Privacy, and Artificial Intelligence: An
Examination of Legal Issues Surrounding the European Union General Data Protection Regulation and
Autonomous Network Defense;’ Volume 21, Issue 1. Retrieved from
https://scholarship.law.umn.edu/cgi/viewcontent.cgi?article=1476&context=mjlst. Published on 6 June, 2020.
3
Mariarosaria Taddeo, Tom McCutcheon, and Luciano Floridi. Nature Machine Intelligence. Trusting artificial
intelligence in cybersecurity is a double-edged sword; volume 1, pages-557–560. Retrieved from
https://doi.org/10.1038/s42256-019-0109-1. Published on 11 November, 2019.
 5

to grow by US$19 billion between 2021 and 2025.4

Despite the positive impact of AI in fortifying cybersecurity, concerns persist. Policymakers,

security practitioners, technologists, researchers, and the public are deeply interested in the
potential implications of AI for cybersecurity. This discussion seeks to highlight the strides
made in applying AI techniques to address cybersecurity challenges, providing an overview
of ethical risks and concerns associated with AI in cybersecurity and suggesting future
directions.

Clarifying Key Concepts: Cybersecurity and AI

While there is currently no universally agreed-upon and standardized definition for both
cybersecurity and artificial intelligence (AI), it becomes crucial to grasp the essence of these
terms, considering their central role in the upcoming discussion.

Deciphering Cybersecurity: Safeguarding Information and Assets

Cybersecurity is fundamentally concerned with safeguarding information and assets,
preventing their misuse by external individuals, while simultaneously ensuring accessibility
for intended users.5 According to Daniel Schatz et al., the concept embodies a strategic
approach and a series of actions tied to security risk management processes implemented by
organizations and states. The overarching goal is to fortify the confidentiality, integrity, and
availability of data and assets within the realm of cyberspace. This comprehensive approach
includes the formulation of guidelines, policies, and a robust collection of safeguards,
technologies, tools, and training.6 Through these measures, cybersecurity aims to provide the
utmost protection for the state of the cyber environment and the well-being of its users.

In essence, cybersecurity is a multifaceted discipline that extends beyond mere technical

measures. It involves a holistic strategy, integrating various elements to counteract potential

4
Ed Bowen, Wendy Frank, Deborah Golden, Michael Morris and Kieran Norton. Deloitte. ‘Cyber AI: Real
defense’. Retrieved from https://www2.deloitte.com/us/en/insights/focus/tech-trends/2022/future-of-
cybersecurity-and-ai.html. Published on 7 December, 2021.
5
Bernd Stahl, Moira Carroll-Mayer, David Elizondo, Kutoma Wakunuma and Yingqin Zheng. SpringerLink.
‘Intelligence Techniques in Computer Security and Forensics: At the Boundaries of Ethics and Law’ in David A
Elizondo, Agusti Solanas and Antoni Martınez-Balleste (eds), Computational Intelligence for Privacy and
Security. Retrieved from https://link.springer.com/chapter/10.1007/978-3-642-25237-2_14. Published on 1
January, 2012.
6
Daniel Schatz, Rabih Bashroush and Julie Wall. The Journal of Digital Forensics, Security and Law. ‘Towards
a More Representative Definition of Cyber Security;’ Volume 12, Number 2. Retrieved from
https://commons.erau.edu/cgi/viewcontent.cgi?article=1476&context=jdfsl. Published on 30 June, 2017.
 6

threats and challenges in the ever-evolving landscape of cyberspace. By understanding and

implementing these measures, organizations and states strive to create a secure digital
environment that upholds the principles of confidentiality, integrity, and availability for data
and assets.

Uncomplicating Artificial Intelligence: A Basic Insight

Artificial Intelligence (AI) can be distilled into computer systems possessing the capacity to
learn from their experiences and tackle intricate problems, resembling the problem-solving
capabilities inherent in human cognition. Alternatively, AI is defined as systems
demonstrating intelligent behavior, autonomously analyzing their surroundings, and taking
actions to fulfill specific objectives to some degree. A fundamental facet of many AI
technologies involves machine learning, a suite of techniques empowering computer
programs to engage in 'thinking' processes by formulating mathematical algorithms based on
accumulated data. The upshot is an algorithm capable of addressing problems or executing
tasks in a manner analogous to the human mind, exemplified in functions like voice or
language recognition. A considerable number of machine learning approaches hinge on deep
learning, a method that artificially emulates neural networks modeled after those present in
the human brain.7

In essence, Artificial Intelligence represents a paradigm where machines mimic human

cognitive processes, evolving beyond mere programmed responses to learn and adapt
dynamically. The incorporation of machine learning, especially through deep learning
techniques, facilitates AI systems in comprehending complex patterns and making
autonomous decisions, furthering their ability to tackle diverse challenges across various
scenarios.

Empowering Cybersecurity Through Artificial Intelligence

In the face of an escalating cyberthreat landscape, the demand for sophisticated tools and
technologies capable of swift detection, investigation, and decision-making in response to
emerging threats has become imperative. Artificial Intelligence (AI) emerges as a potent
7
Jozef Andrasko, Matus Mesarcik and Ondrej Hamulak. SpringerLink. The regulatory intersections between
artificial intelligence, data protection and cyber security: challenges and opportunities for the EU legal
framework. Retrieved from https://link.springer.com/article/10.1007/s00146-020-01125-5. Published on 2
January, 2021.
 7

solution with the potential to intelligently analyze and categorize substantial volumes of
Internet traffic.8 Academic resources highlight the diverse applications of AI techniques in
countering cyber-attacks, with neural networks playing a pivotal role in intrusion detection
and prevention. Beyond this, proposals extend to leveraging neural networks for tasks such as
Denial-of-Service detection, computer worm detection, spam detection, zombie detection,
malware classification, and forensic investigations.

Various AI techniques, including Heuristics, Data Mining, Neural Networks, and Artificial
Immune Systems, find application in advancing next-generation anti-virus technology.
Intelligent agent technology, at times coupled with mobile agent technology, is employed in
some intrusion detection and prevention systems.9 This section provides a concise overview
of existing applications where AI techniques contribute significantly to enhancing
cybersecurity, offering a glimpse into the evolving landscape of AI's role in fortifying digital
defenses.

Exploring the AI Agent Paradigm: Collaborative Intelligence in Action

The AI Agent paradigm, stemming from the notion that knowledge-sharing is essential for
effective problem-solving, represents a distinctive branch of Artificial Intelligence (AI).
Within this framework, a single agent stands as an independent cognitive entity, possessing
its internal decision-making system and a unique goal. To attain its objective, an agent
engages proactively within its environment, collaborating with both the environment itself
and other agents. Notably, agents exhibit reactive behavior, comprehending and responding
to alterations in their surroundings while interacting with both the environment and other
decentralized agents. With the passage of time, these agents autonomously adapt to dynamic
changes in their environments, drawing from their accumulated experiences.10

Given their decentralized and interactive nature, intelligent agents inherently lend themselves

8
Sherali Zeadally, Erwin Adi, Zubair Baig, and Imran A. Khan. IEEE. ‘Harnessing Artificial Intelligence
Capabilities to Improve Cybersecurity’. Retrieved from https://ieeexplore.ieee.org/abstract/document/8963730.
Published on 20 January, 2020.
9
Selma Dilek, Hüseyin Cakır and Mustafa Aydın. ‘Applications of Artificial Intelligence Techniques to
Combating Cyber Crimes: A Review’ (2015) International Journal of Artificial Intelligence & Applications;
Vol. 6, No. 1. Retrieved from https://arxiv.org/ftp/arxiv/papers/1502/1502.03552.pdf. Published on January
2015.
10
Nadine Wirkuttis and Hadas Klein, ‘Artificial Intelligence in Cybersecurity;’ 01 Cyber, Intelligence,
and Security; Volume 1. Retrieved from
https://www.academia.edu/36264684/Artificial_Intelligence_in_Cybersecurity. Published in January 2017.
 8

to gathering comprehensive information across entire networks and surrounding systems.

This characteristic, although employed for defensive purposes, has also found application in
reconnaissance and the exploration of potential target systems. The AI Agent paradigm thus
emerges as a dynamic and adaptive approach, fostering collaboration among entities to not
only defend but also to strategically explore and understand the intricacies of complex
systems.

Brain-Inspired Learning of Artificial Neural Networks (ANNs)

The learning technique employed by Artificial Neural Networks (ANNs) draws inspiration
from the functioning of neurons in the human brain. These techniques represent neurons
through mathematical equations, processing a series of data samples to generate a target
value. Particularly effective in environments where expressing algorithms or rules for
problem-solving proves challenging or where such rules are unknown, ANNs excel at
learning and addressing complex problems. However, due to the somewhat elusive nature of
their system behavior, they are often regarded as undefined black-box models.

ANN techniques showcase remarkable proficiency in discerning patterns within data samples
ranging from noisy to incomplete. This adaptability makes them particularly suitable for
intrusion-detection systems, where they can adjust to novel forms of communication. In a
cybersecurity context, a notable application involves the Cascade Correlation Neural
Network (CCNN), which incrementally introduces new hidden units to the hidden layer as
needed. When new events are detected, the network dynamically incorporates these new
hidden nodes, training them exclusively with the freshly collected data. This results in a
runtime adaptive and scalable system. Another significant advantage of ANNs lies in their
ability to detect zero-day attacks by learning from recent incidents, showcasing their agility
and responsiveness in the ever-evolving landscape of cybersecurity.

Immune-Based Techniques in Cybersecurity

Drawing inspiration from biological immune systems, certain techniques are deployed to
maintain stability in dynamic environments. Immune-based intrusion detection, akin to the
evolution of immunocytes in biological systems, involves simultaneous processes such as
self-tolerance, cloning, and variation. In this context, antibodies are generated by the immune
system to combat pathogens, and the level of intrusion intensity can be gauged by assessing
 9

the concentration of antibodies. This approach, reflective of the immune system's defense
mechanisms, holds significant importance in the realm of cybersecurity, serving as a crucial
defense mechanism against cybercrimes. The implementation of Artificial Immune Systems
(AISs) contributes to the fortification of cyber defenses, aligning with the intricate and
adaptive nature of natural immune systems to safeguard digital environments from potential
threats.11

Expert Systems in AI Applications

Expert systems, integral to the landscape of Artificial Intelligence (AI), are computer
programs meticulously designed to offer decision support for intricate problems within
specific domains, standing as one of the most prevalent applications of AI. Conceptually, an
expert system comprises two primary components: a knowledge base storing expert
knowledge and an inference engine responsible for reasoning about predefined knowledge
and generating solutions for given problems. The application of expert systems varies
depending on the method of reasoning employed, making them versatile across different
problem classes.

One notable approach within expert systems is Case-Based Reasoning (CBR), allowing
problem resolution by recalling similar cases from the past. This method assumes that
solutions from prior cases can be adapted and applied to new problem instances. Continual
improvements occur as newly proposed solutions are evaluated and, if necessary, revised
over time, enhancing accuracy and the system's ability to learn. Rule-Based Systems (RBS)
form another facet, solving problems using rules defined by experts, encompassing a
condition and an associated action. Unlike CBR systems, RBSs lack the capacity to
autonomously learn new rules or modify existing ones, leading to the "knowledge acquisition
problem," particularly significant in dynamic environments. In the realm of cybersecurity,
expert systems find widespread use among security professionals, offering invaluable
decision support in navigating complex cyber environments.

Leveraging AI Advantages in Cybersecurity

Artificial Intelligence (AI) emerges as a transformative force in cybersecurity, offering a

11
Zeeshan Siddiqui, Sonali Yadav and Mohd Shahid Husain. ‘Application of Artificial Intelligence in Fighting
against Cyber Crimes: A Review;’ 9 International Journal of Advanced Research in Computer Science
118. Retrieved from file:///C:/Users/Admin/Downloads/aceit18.pdf. Published in April, 2018.
 10

range of advantages to combat the ever-evolving landscape of cyber threats. The dynamic
nature of cyberattacks and the proliferation of devices necessitate a response that AI can
provide more effectively than traditional cybersecurity methods. Here are several key
advantages of applying AI in the realm of preventing cyber threats:

1. Large Data Processing: AI enables organizations to process vast amounts of data with
precision and efficiency.12 The ability to scan extensive data sets and utilize Machine
Learning (ML) algorithms to discern patterns empowers AI to guide human decision-
makers on effective response strategies. The speed and scale at which AI software
accomplishes data analysis tasks surpass the capabilities of human security experts.13
2. Early Detection of New Risks: Traditional software struggles to keep pace with the
volume of new malware created by attackers.14 AI-powered cybersecurity solutions
excel in identifying both novel, unknown threats and known threats that have
breached networks. This capability plays a crucial role in mitigating advanced
hacking techniques employed by attackers.15
3. Fast Response: AI significantly reduces the time taken to detect threats and breaches,
enhancing overall cybersecurity. The use of AI can cut down the time for remediation
and patch implementation by up to 12%. Some organizations achieve even greater
reductions, exceeding 15%, contributing to a more agile response to cyber threats.
4. Higher Accuracy: AI-based cybersecurity data analysis consistently outperforms
human analysts in accuracy, especially in large-scale data analysis and anomaly
detection. AI excels in reducing errors and faults in operational tasks, showcasing its
efficiency in finding anomalies and irregularities.
5. Cost Reduction: AI-driven cybersecurity systems comprehend and reuse threat
patterns, leading to a reduction in the time and effort required to identify, investigate,
12
Ljubomir Lazic. ‘Benefit From AI in Cybersecurity;’ The 11th International Conference on Business
Information Security, Belgrade, Serbia. Retrieved from
file:///C:/Users/Admin/Downloads/Prof.Lazicpaper_BISEC2019Ver.2.pdf. Published in Oct, 2019.
13
Watney, M M. Artificial Intelligence and its' Legal Risk to Cybersecurity; Originally published in
Proceedings of the 19th European conference on cyber warfare and security (Academic Conferences and
publishing International limited, UK 2020). Retrieved from
https://www.proquest.com/openview/ce8ef72f9d9548ffeb1dcf2e667eba03/1?pq-
origsite=gscholar&cbl=396497#:~:text=AI%2Ddriven%20cybersecurity%20technology%20must,should%20be
%20applied%20to%20cybersecurity. Published in June 2020.
14
Gaurav Belani. The Use of Artificial Intelligence in Cybersecurity: A Review. Retrieved from
https://www.computer.org/publications/tech-news/trends/the-use-of-artificial-intelligence-in-cybersecurity
Published on 01/16/2021.
15
Kashyap Vyas. IT Business Edge. ‘The Pros and Cons of Enlisting AI for Cybersecurity’. Retrieved from
https://www.itbusinessedge.com/security/the-pros-and-cons-of-enlisting-ai-for-cybersecurity/. Published on 22
October 2021.
 11

and remediate threats. A majority of executives report that AI lowers the cost of
detecting and responding to breaches, with reductions ranging from 1% to 15%, and
some organizations achieving even higher cost reductions.
6. Filling Employment Gap: The shortage of cybersecurity talent worldwide is
exacerbated by the escalating severity and complexity of cyber threats. AI plays a
pivotal role in filling this employment gap, offering a strategic solution to the chronic
shortage by automating certain tasks and augmenting the capabilities of existing
cybersecurity professionals.

AI Drawbacks in Cybersecurity
Despite the promising advancements in AI for cybersecurity, certain drawbacks need
consideration:

1. Asymmetry: Defenders often face a significant disadvantage in terms of timing,

knowledge, and preparation compared to attackers. Defenders must maintain a
constant high-level defense against unknown adversaries, facing challenges in
deploying defensive technologies promptly.
2. Limited Datasets: The practice of leveraging single datasets in a siloed manner limits
the comprehensiveness of cybersecurity appraisals. Access to multiple datasets and a
deeper understanding of their interrelationships are crucial for a more holistic
understanding of the cyber environment.16
3. Lack of Model Explainability: Prevailing AI-based algorithms in cybersecurity,
particularly those relying on deep learning, are often characterized by a "black-box"
nature, lacking model explainability and interpretability. This hinders trustworthiness,
security, and privacy, crucial elements in cybersecurity applications.17
4. Human in Loop: Despite significant advancements, AI-based security systems are not
fully autonomous. Human intervention is still necessary for certain tasks, emphasizing
the need for a close linkage between AI-based approaches and human actions to
enhance decision-making processes.

16
Sagar Samtani, Murat Kantarcioglu and Hsinchun Chen. Trailblazing the Artificial Intelligence for
Cybersecurity Discipline: A Multi-Disciplinary Research Roadmap (Volume 11; Issue 4; Article No.: 17; pp 1–
19). Retrieved from https://dl.acm.org/doi/pdf/10.1145/3430360. Published on 2 December 2020.
17
Ibid.
 12

While AI holds great potential in fortifying cybersecurity, addressing these drawbacks is vital
to ensure effective, transparent, and collaborative cyber defense strategies.

New Frontiers of Cyber-Attacks on AI Systems

Traditionally, cyber-attacks primarily aimed at data theft and system disruption. However, the
emergence of new attack forms specifically targets AI systems, seeking to gain control and
manipulate their behavior, thereby undermining the potential improvements offered by AI.
These attacks exploit the learning capabilities of AI systems, with three notable types:

1. Data Poisoning: Attackers inject carefully crafted erroneous data into the legitimate
training data, altering the behavior of the AI system.
2. Tempering of Categorization Models: By manipulating the categorization models of
neural networks, attackers can deceive AI systems into misclassifying objects, as
demonstrated by the manipulation of 3D-printed turtle images, causing an AI system
to misclassify turtles as rifles.
3. Backdoor-Based Attacks: These attacks rely on hidden triggers added to the AI
model, overriding correct classifications and making the system perform
unexpectedly. For instance, a study added images of stop signs with a special sticker
to the training set, causing an AI system to classify any stop sign with that sticker as a
speed limit sign, posing significant safety risks for autonomous vehicles.

The Ethical Quandaries of AI in Cybersecurity

The integration of AI technologies into cybersecurity, while powerful, raises ethical concerns
that often escape the understanding of ordinary users. Several ethical problems stem from the
application of AI in cybersecurity:

 Data Privacy Concerns: AI, particularly advanced techniques like Artificial Neural
Networks (ANNs), poses challenges to data privacy. The extensive analysis of data,
including files, emails, and user behavior, for threat detection can lead to privacy
concerns. The continuous monitoring required for enhancing system resilience raises
questions about user privacy, exposes them to additional risks in case of data
breaches, and may contribute to a mass surveillance effect.18
18
Mariarosaria Taddeo, ‘Three Ethical Challenges of Applications of Artificial Intelligence in Cybersecurity’
 13

 Surveillance Impact: Beyond privacy, AI-driven surveillance has psychological and

social consequences, potentially resulting in social classification, stigmatization, and
exclusion. The extensive data collection and analysis for AI threat identification may
contribute to surveillance practices, impacting society in ways that affected
individuals may not be aware of.
 High Automation Risks: The high level of automation in AI systems, discovering
rules during training or operation, poses risks by potentially reducing human
situational awareness19. While automation can replace repetitive tasks, it may also
remove subtle discretionary decision-making, impacting the immediate operation of
systems and human involvement.
 AI Bias and Discrimination: AI systems, like humans, can exhibit biases and
discriminate based on factors such as gender, race, age, disability, and ideology. 20
Biased AI, as observed in criminology prediction systems and facial recognition
studies, can perpetuate and even exacerbate existing biases present in the datasets
used for training.21
 Shift in Power Relationships: Ownership of powerful AI security technologies has the
potential to alter existing power relationships. Employers, utilizing AI for network
security, could misuse these tools to monitor and detect activities they deem
undesirable among employees, raising questions about the ethical grounds, and
purposes for which such technologies are used.

Addressing these ethical challenges requires a thoughtful and responsible approach to the
development, deployment, and use of AI technologies in cybersecurity, ensuring
transparency, fairness, and adherence to ethical principles.

(2019) 29 Minds and Machines 187. Retrieved from https://doi.org/10.1007/s11023-019-09504-8. Published on

03 June 2019
19
Joshua A. Kroll, James Bret Michael and David B. Thaw. Enhancing Cybersecurity via Artificial Intelligence:
Risks, Rewards, and Frameworks. Retrieved from
https://www.dtra.mil/Portals/125/Documents/Media-Kit/Kroll-June2021.pdf. Published on 7 June 2021.
20
Jasmine Henry. Biased AI Is Another Sign We Need to Solve the Cybersecurity Diversity Problem. Retrieved
from https://securityintelligence.com/articles/biased-ai-is-another-sign-we-need-to-solve-the-cybersecurity-
diversity-problem/. Published on February 6, 2020.
21
Hussein, Burhan, Haleem, Chongomweru and Siddique, Muhammad. The Future of Artificial Intelligence and
its Social, Economic and Ethical Consequences. Retrieved from
https://www.researchgate.net/publication/
348403835_The_Future_of_Artificial_Intelligence_and_its_Social_Economic_and_Ethical_Consequences.
Published on 2021/01/09.
 14

Conclusion
As the integration of AI systems in cybersecurity accelerates, it is imperative to subject them
to ethical scrutiny to mitigate risks such as deskilling and unwarranted surveillance that may
arise from uncontrolled behavior. Faced with potential risks, society may contemplate
rejecting the application of AI techniques in cybersecurity, despite their capacity to enhance
data system security. However, limited adoption could render cyber defense more vulnerable,
providing malicious users an advantage in crafting novel attacks that exploit the absence of
robust AI defenses.

The initial crucial step involves designing and deploying AI in strict adherence to ethical
standards, ensuring responsible and accountable behavior. The establishment of AI
regulations becomes paramount to uphold ethical norms and safeguard individual rights.
Moreover, through collaborative efforts spanning civil society, politics, business, and
academia, coordinated strategies can be identified and pursued. This collaborative approach
aims to harness the potential of AI as a force for human flourishing while upholding the
fundamental principle of respecting human dignity. By navigating these ethical pathways, AI
in cybersecurity can evolve into a beneficial and protective force that aligns with the well-
being of individuals and society at large.