INTERNATIONAL JOURNAL OF SCIENTIFIC RESEARCHERS
ISSN: 3030-332X Impact factor: 8,293
Volume 12, issue 1, June 2025
https://wordlyknowledge.uz/index.php/IJSR
worldly knowledge
Index:
google scholar, research gate, research bib, zenodo, open aire.
https://scholar.google.com/scholar?hl=ru&as_sdt=0%2C5&q=wosjournals.com&btnG
https://www.researchgate.net/profile/Worldly-Knowledge
https://journalseeker.researchbib.com/view/issn/3030-332X
119
CYBER SECURITY IN THE AGE OF ARTIFICIAL INTELLIGENCE
Djumakulova Shaxlo Davlyatovna,
Quyliyev Sardorbek Abdixoliq ugli
Computer science teachers of the Academic Lyceum of
Termez State University of Engineering and Agrotechnology
Annotation
. This article explores the evolving relationship between artificial intelligence and
cybersecurity. It explains how AI improves threat detection, response automation, and
predictive analysis, while also introducing new risks such as adversarial attacks, data poisoning,
deepfakes, and AI-powered phishing. The text emphasizes the need for ethical guidelines,
human oversight, and international cooperation to maintain a secure digital environment in the
age of intelligent technologies.
Keywords
: artificial intelligence, cybersecurity, adversarial attacks, phishing, deepfake, data
poisoning, threat detection, machine learning, cyber threats, AI security, automation, ethical AI,
digital privacy, international regulation, SI-based attacks.
Artificial Intelligence (AI) has become an important tool in today's digital security
challenges. At the same time, SI technologies themselves are creating new types of risks. The
relationship between cybersecurity and artificial intelligence is complex: on the one hand, the
fight against cyberattacks is increasing with the help of AI, and on the other, hackers are using
AI tools to create more advanced threats. This has fundamentally changed the modern security
environment. First of all, AI has increased the ability to automatically detect threats, analyze
them in real time, and prevent them. For example, machine learning algorithms help detect
cyberattacks at an early stage by detecting unusual behavior in the network. Unlike traditional
security systems, AI is constantly learning, improving itself, and adapting to new types of
threats.
However, this very self-learning feature also creates new vulnerabilities in AI-based
security systems. For example, hackers can deliberately misdirect an AI system through so-
called “adversarial attacks.” These attacks use specially designed data sets to confuse the AI
model, causing the system to make incorrect decisions. This is especially a threat to
systems that provide facial recognition, access control, or financial security. In addition,
hackers themselves have begun to use AI tools. For example, automated phishing campaigns
and social media-adapted phishing emails are becoming more credible with AI. This makes it
more difficult for ordinary users to detect and eliminate the threat. Disinformation campaigns,
“deepfake” videos, and voice spoofing using AI are also new challenges in cybersecurity. Such
tools can be used to commit crimes such as political manipulation, defamation of an individual
or organization, and fraud.
The database (dataset), which is an important part of the operation of artificial
intelligence, can also become a weak point. If this data is manipulated, the AI will make
the wrong decision. Therefore, a type of threat called “data poisoning” requires serious
INTERNATIONAL JOURNAL OF SCIENTIFIC RESEARCHERS
ISSN: 3030-332X Impact factor: 8,293
Volume 12, issue 1, June 2025
https://wordlyknowledge.uz/index.php/IJSR
worldly knowledge
Index:
google scholar, research gate, research bib, zenodo, open aire.
https://scholar.google.com/scholar?hl=ru&as_sdt=0%2C5&q=wosjournals.com&btnG
https://www.researchgate.net/profile/Worldly-Knowledge
https://journalseeker.researchbib.com/view/issn/3030-332X
120
attention in the field of cybersecurity. Also, the excessive autonomy of AI-based security
systems poses a problem. For example, if an AI-based security system automatically blocks
access to the system or marks the wrong user as “dangerous”, an environment where
technology controls people will arise. This raises serious questions about ethics and human
rights. At the international level, the issue of regulating cybersecurity operations based on
artificial intelligence has not been resolved. Different countries are pursuing their own policies
in this regard, and these technologies are becoming an integral part of national security. This
indicates the need for global agreements and ethical standards.
In the future, defense systems developed with the help of AI will be even more perfect:
they will be able to predict hacking attacks in advance, immediately assess the risk and take
specific measures. But human oversight, ethical constraints, transparent algorithms, and
international standards play a crucial role. Otherwise, AI-based cybersecurity systems could
become self-regulating, closed systems that do not require human intervention, putting the
information society at risk. The evolution of artificial intelligence in cybersecurity has not only
been a technological revolution, but also a paradigm shift in security. Now, no digital system
can provide robust security without the help of AI. However, the transformation of this
technology into an effective weapon against cyberattacks, in turn, makes it a target.
One of the main principles of cybersecurity, "proactive defense", is now implemented
through AI. Previously, security systems responded only after an attack occurred. Now, AI
allows you to predict threats, analyze behavior and take countermeasures before damage occurs.
This is especially important for the financial sector, banking system, healthcare systems and
government infrastructure. For example, "SIEM" (Security Information and Event Management)
systems developed based on artificial intelligence analyze millions of log records in the
network in real time and identify signals of potential attacks. They detect deviations from
normal user behavior and alert the administrator or take automatic action. Also, “User and
Entity Behavior Analytics” (UEBA) technologies use AI algorithms to track user behavior and
detect malicious activity. AI defines this as a threat if a user logs in at a different time than
usual, connects from an unknown device, or downloads a large amount of data.
However, despite these capabilities, AI technologies have their own vulnerabilities.
Under the term AI Supply Chain Attacks, hackers are now attacking the models used to create
AI algorithms or their databases. They can add malicious data to the model training process and
confuse it. As a result, AI-based security systems begin to malfunction. Another dangerous
situation is the proliferation of AI as a Service (AIaaS) platforms. Through these platforms, any
user can access AI services. Hackers can use such services to plan attacks, test them, and detect
“zero-day” vulnerabilities. This makes artificial intelligence-based attacks popular and
affordable.
References
:
1. Russell, S., & Norvig, P. (2020). Artificial Intelligence: A Modern Approach (4th ed.).
Pearson.
2. Brundage, M., et al. (2018). The Malicious Use of Artificial Intelligence: Forecasting,
Prevention, and Mitigation.
3. Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep Learning. MIT Press.
