Page 1 of 10

Transactions on Engineering and Computing Sciences - Vol. 13, No. 02

Publication Date: April 25, 2025

DOI:10.14738/tecs.1302.18411.

Arvin, S., Hendricks, N., & Ketel, M. (2025). Artificial Intelligence, Cybersecurity, and a Growing Ethical Dilemma. Transactions on

Engineering and Computing Sciences, 13(02). 19-28.

Services for Science and Education – United Kingdom

Artificial Intelligence, Cybersecurity, and a Growing Ethical

Dilemma

Steven Arvin

Applied Information Technology Department,

University of Baltimore, Baltimore, MD 21201, USA

Nigel Hendricks

Applied Information Technology Department

University of Baltimore, Baltimore, MD 21201, USA

Mohammed Ketel

Applied Information Technology Department

University of Baltimore, Baltimore, MD 21201, USA

ABSTRACT

Currently, cybersecurity threats, particularly cyber-attacks, are a growing concern.

As time goes on, it becomes increasingly challenging to hinder these attacks.

Nevertheless, a new participant has entered the arena, known as artificial

intelligence (AI). AI offers a way for cybersecurity experts to counteract the ever- evolving attacks. By utilizing techniques such as identifying threats and automated

responses to incidents, organizations can enhance their security measures and

safeguard confidential data. Despite the numerous advantages of adopting AI, it is

equally important to remain vigilant about potential risks. In recent years, the rapid

growth in cybersecurity threats has necessitated the development of more effective

measures to protect sensitive information and systems. This paper delves into the

ethical concerns of AI in cybersecurity, stressing the crucial balance between

technological innovation and maintaining ethical standards.

Keywords: Cybersecurity, Artificial Intelligence, Machine Learning, Deep Learning, IoT,

Ethics.

INTRODUCTION

Artificial Intelligence may seem like an overused buzzword to many, but its utility and

application are growing with each passing day. Whether the logic component of an artificial

intelligence opponent in online chess or the extra set of eyes to see trends in a financial market

that a human would miss, Artificial Intelligence is everywhere. With this ground-breaking field

advancing at a breakneck pace, there are drawbacks as well. Often, technology advances

significantly faster than regulations and laws permit. This can result in business and personal

practices that are ethically questionable at best and dangerous at worst. Sometimes, the best

of intentions can come with negative consequences. This can be seen in the world of

cybersecurity. Artificial Intelligence is readily being applied to cybersecurity systems to

provide active threat avoidance and defense. However, Artificial Intelligence is also being

utilized by bad actors to attack vulnerable systems and victims. Like all things, it takes a

Page 2 of 10

20

Transactions on Engineering and Computing Sciences (TECS) Vol 13, Issue 02, April - 2025

Services for Science and Education – United Kingdom

balancing act, accepting the good with the bad. But first, it helps to have a grasp on what

Artificial Intelligence and cybersecurity are, along with how they can interact [2].

However, even with a firm grasp on the capabilities, pitfalls, and dangers of AI applications,

there is only so much a person can be expected to prepare for. In addition, a private citizen is

unlikely to have the knowledge, access, or resources to procure and implement hardware,

software, and experience into a personal list of best practices. These private citizens are

susceptible to an online attack now more than any other time in history between the ubiquitous

implementation of Internet of Things devices and the decision by many industries to move

services to an online platform. Also, history hasshown that businesses will not always

implement practices that protect the consumer and can even be found to take advantage of

vulnerable parties. For this reason, large governmental organizations, such as the European

Union and the United States Federal Government have begun to layout guiding principles for

how both public and private entities will be expected to utilize AI for the safety of all parties

involved. This is a step in the right direction, but time will tell where more emphasis will need

to be placed. [11, 12].

UNDERSTANDING ARTIFICIAL INTELLIGENCE

In order to understand how Artificial Intelligence (AI) and Machine Learning (ML) can be

applied to cybersecurity safeguards and systems, it is first important to understand what AI is

and how it operates. While in the not-too-distant past, AI was reserved for science fiction media

and people with wild imaginations, it has now permeated many facets of society. It can be found

in the malicious mail filters on email servers, online customer service chatbots, and the

backbone of the logic component in computer games, as well as video games. While these

applications of AI may seem trivial or marginally beneficial to many, they are just a small

glimpse into its capabilities and a foundation for the vast possibilities in its future [8].

For many day-to-day activities and applications, traditional programming can be applied to

satisfy a requirement. This can be covered by utilizing fuzzy sets, data structures, algorithms,

and rules. Unfortunately, these traditional programming methods are not all-encompassing

though. There will always be scenarios where a traditional program is presented with a

problem that it is unable to solve via the resources available to it. This is where AI can bridge

the disconnect or shortcoming. For example, since AI and ML are, in most cases, an extension

of human intelligence, their learning capabilities and analysis of historical trends can be applied

to financial models [8]. While data on users as a significant driver of revenue for large online

conglomerates such as Meta, Google, and others is a relatively new phenomenon, the need to

mine financial data is a long-held premise. The trends of markets, goods, and Gross Domestic

Product (GDP) are arduously studied by analysts globally, looking for any competitive

advantage that can be leveraged. However, they are limited to the models available to them

and the inferences they can make from them based on what they are taught, their personal

experiences, and what logical conclusions they can draw themselves. AI and ML offer an

additional avenue of insight. They can draw conclusions that would seem illogical to a

traditionally trained analyst, but historically, see a pattern or trend that could generate or save

a bank significant sums of money [2].

Page 3 of 10

21

Arvin, S., Hendricks, N., & Ketel, M. (2025). Artificial Intelligence, Cybersecurity, and a Growing Ethical Dilemma. Transactions on Engineering and

Computing Sciences, 13(02). 19-28.

URL: http://dx.doi.org/10.14738/tecs.1302.18411

LEVERAGING MACHINE LEARNING TO ENHANCE CYBERSECURITY

Machine Learning at Glance

As stated above, ML is a quintessential component of AI as a whole. ML is the term for a

computer or machine’s ability to “learn”. While these devices do not have consciousness like

humans, they are able to sort through vast amounts of data and draw conclusions, as well as

see patterns. This is an incredibly simplistic description of how ML occurs. An example of ML

at work would be a program that reads handwriting. The best way to “teach” the program

would be to feed it as many handwriting samples as possible. This can be done manually via

sample collection, or a simple web scraping program can be developed to feed the program

millions of handwriting samples, readily available for free online. The ML program will then

digest these samples and draw a long list of conclusions that it will utilize when reading future

handwriting. Some of the conclusions or rules the ML program may develop are the differences

between print and cursive, what characters are letters, numbers, or symbols and then it could

even begin to realize when punctuation is missing or used incorrectly. Over time, the program

will continue to digest more and more data, making it better equipped to analyze what it is

reading. The goal is that the longer this program is operational and refined, the better it will be

at its job [8].

An Overview of Cybersecurity

While the term Cybersecurity is used on a near daily basis, in regular conversation, and from

the news media, many may not appreciate what it actually means and what it entails. At its

core, cybersecurity is every policy, procedure, software, application, and hardware interface,

that are utilized to protect the integrity of a network and all of the information that is contained

and transmitted throughout the host network. This can be comprised of firewalls, network

intrusion systems, policies that outline what type of web traffic is permitted behind a network

firewall, along with the permitted variety ofoutgoing web traffic and a host of other products

and standard operating procedures. It is also important to appreciate the reason for the ever- increasing emphasis on cybersecurity. As technology and the internet continue to bleed into

every facet of daily life, so does the need for data and financial protection, which all fall under

the umbrella of cybersecurity. Analysts estimate that the average data breach costs an entity

just shy of 4 million dollars to recover from and when the data breach is specifically directed at

a company based in the United States, it climbs to over 8 million dollars. While these numbers

alone should be enough to give many corporate executives pause when assessing their

cybersecurity budgets, it is arguably more important to consider the personal user data

collected and stored by many companies. Private financial data, medical records and personally

identifiable information (PII), such as social security numbers are frequently collected by

companies and stored for tax and identification purposes. If a company is unable to show they

made an earnest attempt to protect this user data, they will lose the trust of their clients and

potentially find themselves at the receiving end of a civil or even criminal court case.

Thankfully, AI solutions can be implemented to help combat these breaches [9].

ML at Work in Cybersecurity

As mentioned earlier referenced, ML plays an invaluable role in AI implementation in

cybersecurity systems. ML programs are developed over time by parsing and digesting large

amounts of data. This can be as simple as tracking which ports are most frequently used by an