Volume-3, Issue-5, May 2025 International Journal of Modern Science and Research Technology

ISSN No- 2584-2706

A Study on Impacts of Artificial Intelligence in

Cyber Security
Sandhya Dahake ; Mayuri Rangari
Manthan Thaware ; Yash Waghmare
Department of MCA G H Raisoni College of Engineering and Management,
Nagpur, Maharashtra, India.

Abstract
This paper examines the impact of AI in Cyber In the digital age, cyber security has become an
Security, Challenges of AI in cyber security. important concern for individuals, businesses
Despite the benefits, using AI in cyber security and authorities worldwide. Increasing
makes the challenges: Adversarial AI attacks - Sophistication of cyber threats - ranging from
can manipulate cyber criminals AI models to harmful software and fish attacks to advanced
detect or prejudice in the security algorithms. A Consistent threats (APTS) - made traditional
Study on impacts of AI in Cyber-security security methods less effective. As cyber
Ethical thought-of-driven security systems criminals utilize digital infrastructure
should balance automation with moral weaknesses, the need for adaptive and
concerns such as users' privacy and algorithm intelligent security solutions has increased
bias. Data dependence and accuracy-AI significantly.
security models require high-quality extensive Artificial intelligence (AI) has proven to be a
data sets to detect effective threats, and raise transformation force in cyber security, to detect
concerns about data security and reliability. danger, to automate security reactions and
The distribution of regulatory and legal barriers provide future indicative analysis to develop
AI-controlled cyber security solutions should cyber risk. AI operated cyber security system
follow the legal framework to ensure moral can analyse large amounts of data, identify
implementation. The importance of AI in cyber non-conformities and adapt to real-time rescue,
security. Traditional cyber security models offer an active approach to digital security.
depend on rule -based systems and manual The importance of AI in cyber security.
threat analysis, which can be time -consuming Traditional cyber security models depend on
and ineffective against quickly developed rule -based detection systems and manual
cyber-attacks. AI provides automation and threat analysis, often slowly and disables to
intelligence information for security handle large -scale cyber threats effectively.
operations, which allows the system: Analyse However, AI introduces automation and
the huge versions of real -time safety data. Find intelligence in security operations, enables:
and reduce cyber dangers with minimal human Detection and mitigation of the risk of real time
intervention. Learn from previous attacks and when using machine learning algorithms.
customize safety strategies accordingly. Behavioural analysis to identify suspicious
Increase user approval through biometric AI activities. Automatic safety event with minimal
system. Media cyber-lock pattern using human intervention. Future analysis to remove
behavioural analysis. potential cyber risk before becoming
physical.AI-enhanced biometric authentication
Keywords: Artificial Intelligence (AI), Cyber for better user verification. By integrating AI
security, Machine Learning, Threat Detection, into cyber security structure, organizations can
AI-driven Security, Cyber Threats. increase their ability to handle zero-day
attacks, inside hazards and advanced Cyber
1. Introduction espionage strategy.

IJMSRT25MAY045 www.ijmsrt.com 145

DOI: https://doi.org/10.5281/zenodo.15462132
Volume-3, Issue-5, May 2025 International Journal of Modern Science and Research Technology
ISSN No- 2584-2706
2. Literature Review technology to increase the safety and
1. Introduction to AI in Cyber security prevention of fraud.
Artificial intelligence (AI) has proven to be a  AI-driven future analysis-AI model to
transformation force in cyber security, predict cyber threats before continuing.
expanded the danger, increases automated  Ethical AI development -strengthen the AI
safety reactions and future indicative analysis. regime to ensure fair, fair and transparent
AI-driven safety models benefit from machine security models.
learning, neural networks and natural language
treatment (NLP) to identify cyber threats and 3. Methodology
reduce the risk. Research Approach: This study uses a
systematic research method, which integrates
2. Existing Research on AI in Cyber security both qualitative and quantitative analysis to
Many studies have discovered the role of AI evaluate the effect of AI in cyber security. The
in cyber security, and postpone both benefits feature is designed to detect AI-operated safety
and challenges: mechanisms, their applications, challenges and
 The AI-controlled Far detection AI improves future implications.
safety by analysing giant datasets to detect
anomalies and predict cyber-attacks. A qualitative approach can analyse the study
 Adversarial AI Trimmer - Research and trends in the case, while a quantitative
discusses how cyber -criminal AI is utilized approach can evaluate matrix such as
to circumvent security, which emphasizes improving the speed or efficiency of AI-
the need for adaptive AI -security models. operated units in Cybercrime.
 AI-based scam prevention-of-manual Highlight the logic for choosing this specific
behavioural analysis improves fraud approach, and emphasizes how it matches
detection and identification of deviations in research goals.
financial transactions.
 Ethical thought-studying AI-operated cyber Data Collection Methods: Primary data:
security emphasizes the concerns of privacy, Mention basic data you collected, such as
algorithm bias and openness. examination or expert interview about the use
of AI in cyber security.
3. Challenges and Limitations in AI Cyber
security Secondary data: Explain the use of already
 Despite the progress of AI, researchers have existing data sets, academic magazines, white
identified many challenges: articles and industry reports from reliable
 Data dependency-AI security models require sources such as NIST, IEEE or large cyber
high-quality data sets, which raises concerns security firms.
about data protection and accuracy. Clearly describe your process of purchasing,
 Regulatory compliance-of-operated cyber organization and curating relevant information
security should be adapted to legal for the study.
frameworks to ensure moral
implementation. Analytical Framework:
 AI-operated cyber-attack criminals use rapid Qualitative Analysis: Appreciation Analysis to
AI to develop refined attacks, requiring identify patterns, concepts and conditions
continuous AI-security reforms. related to AI in cyber security.
Quantitative Analysis: Statistical analysis to
4. Future Directions in AI Cyber security measure factors such as response time, AI-
Research based success rate for cyber security equipment
 AI-Block chain Integration-Data and frequency of AI-related cyber-attacks.
combination of AI with block chain
Validation Methods:

IJMSRT25MAY045 www.ijmsrt.com 146

DOI: https://doi.org/10.5281/zenodo.15462132
Volume-3, Issue-5, May 2025 International Journal of Modern Science and Research Technology
ISSN No- 2584-2706
Triangulation: Check conclusions using many Estimated AI adoption in 2024: The graph
data sources or methods. indicates that AI -adoption in cyber security
Peer Review: Mention if any conclusion was has crossed 70% by 2024. This reflects the
reviewed or validated by experts in AI or cyber widespread use of AI in the danger that detects,
security sectors. improves the response time and increases the
Reliability: Other researchers highlight the overall cyber security defence.
steps taken to repeat the study and ensure
frequent results. 4. Benefits
 Enhanced Detection Capabilities: AI
Limitations: accept any challenges or obstacles algorithm detects frequent dangers (APTS)
to validating data, such as limits in prejudice or and unknown malware variants by advanced
test scenarios in secondary sources. advancement than signature-based
approaches [9].
 Speed and Automation: AI, while reducing
human effort and response time, automating
Far detectionand response [12]. A real -time
threat information ensures rapid murdering
security breach.
 Better accuracy: AI-controlled system
reduces false positivity and false negative in
threat detection [8]. The ability to
distinguish between real threats and benign
activities helps safety teams to focus on
important issues.
 Scalability: AI activates cyber security
solutions effectively with growing digital
environment [16]. AI-operated clouds help
Interpretation of diagram: Adoption of AI in
security solutions organizations secure their
cyber security throughout the year the chart
expansion of digital infrastructure without
that comes with reflects the growth of AI in
compromising on performance
cyber security from 2015 to 2024. This graph
 Cost reduction: AI-driven cyber security
includes Store Takeout’s: Stable increase in AI
manual detects danger and reduces costs
adoption: In 2015, the AI adoption in cyber
associated with response, so that
security was less than 10%. By showing a
organizations can distribute resources more
positive tendency, the adoption rate
efficiently [17].
continuously increased every year.
Fast growth post -2018: Between 2018 and
5. Challenges
2020, there was a significant increase in the AI
Side effects: Cyber criminals inject misleading
adoption, showing that organizations began to
data and manipulating the AI model and
recognize the efficiency of the AI-controlled
causing spontaneous abortion of hazards. This
security solutions. AI-operated Faring
is a great concern as the attackers develop
Detection, Infiltration Detection System (IDS)
unfavourable strategies to bypass the AI-
and Malware analysis gained traction in cyber
operated safety mechanism.
security structure.
Prejudice in AI model: AI system training can
Fast growth after 2020: The AI decision rate
receive bias from the dataset, which can lead to
exceeded 40% by 2021 and increased rapidly.
incorrect or incorrect hazard assessment. The
This suggests a change in industry practice, as
bias in AI can lead to false positivity or
AI became a basic component of cyber security
negative, affecting the effectiveness of the
strategies.
rescue of cyber security.

IJMSRT25MAY045 www.ijmsrt.com 147

DOI: https://doi.org/10.5281/zenodo.15462132
Volume-3, Issue-5, May 2025 International Journal of Modern Science and Research Technology
ISSN No- 2584-2706
Data Privacy and Compliance: AI-based cyber Fish prevention: AI-driven fish declaration
security solutions require extensive data systems use natural language treatment (NLP)
collection, increasing the concerns of the user's to analyse e-mail content, URL and sender
privacy and regulatory compliance, such as behaviour, have high accuracy when it comes
GDPR and CCPA requirements. to detecting fish attacks (Bahnsen et al., 2017)
High calculation costs: Distributing AI- [1].
operated security solutions requires important Danger Intelligence and prediction: The
calculation resources, which leads to expensive process of large amounts of danger of
to use AI-based cyber safety equipment for identifying the pattern of AI attacks makes the
small outfits. process of intelligence data and predicts
Lack of clarity: Many AI models act as black potential cyber threats before it (McAfee Labs,
boxes, making it difficult for security analysts 2020) [4].
to explain how decisions are made. This lack of Automatic event reaction: AI-controlled
openness can prevent trust and responsibility in security automation quickly separates the
AI-operated security solutions. dangers and reduces the response time in cyber
False positivity and negative: AI-based security security incidents, highlighted by Sharmen et
system sometimes generates misinformation, al. (2021) [5].
either flags the mild activities that dangers Scalability and cost efficiency: AI-driven
(false positive) or fail to detect real threats security solution allows organizations to be
(false negatives), leading to safety intervals. effectively cyber security skills, which reduces
Ethical concerns: The use of AI in cyber manual efforts and operating costs [16, 17].
security improves moral concerns about Explanation for openness AI (XAI): AI models
monitoring, computer ownership and potential develop to become more interpretable, so that
abuse of AI-operated safety equipment. security teams can understand decision -
Emerging quantum threat: In the form of making, reduce prejudice and improve
progress of quantum calculation, current AI- reliability [11].
operated encryption techniques can be unsafe, Integration with block chain for secure
requiring a new safety approach to combat authentication: A combination of AI and block
potential risks. chain increases data protection and certification
processes, making them tampering [15].
6. Features of Ai in Cyber Security Federated Learning (FL) Learning for Privacy
Automatic Far Disconnection: AI models can Conservation Safety: AI models transfer to
analyse real -time network traffic and detect federated learning, which allows organizations
cyber threats with minimal human intervention. to improve cyber security without sharing
AI-enhanced intrusion detection systems (IDS) sensitive data [17].
improve attacks on zero day (Buczak & Guven,
2016) [3] improve the skills of identifying 7. Conclusion
significantly. The integration of AI into cyber security has
Detection of behavioural analysis and increased the danger, reaction efficiency and
deviation: AI monitors user activity and system general safety management. AI-controlled
behaviour to detect anomalies that may indicate solutions, such as infiltration detection systems
cyber threats. Behaviour-based AI solutions and fish declaration models, show high
reduce false positivity compared to traditional accuracy, especially when combined with NLP
rule-based methods [10]. techniques and real-time monitoring.
Detection of advanced harmful software: AI Automatic AI-driven reactions reduce the event
utilizes deep learning to identify and classify reaction time and offer an active approach to
harmful software with high accuracy. The AI reducing cyber risk. However, challenges
models consistently learn to improve the remain, including data set quality, model
detection features of new Malware variants adaptation and moral concerns.
[13].

IJMSRT25MAY045 www.ijmsrt.com 148

DOI: https://doi.org/10.5281/zenodo.15462132
Volume-3, Issue-5, May 2025 International Journal of Modern Science and Research Technology
ISSN No- 2584-2706
Explanation for openness AI (XAI): AI models [12] Lippmann, R. P., et al. (2000). "Evaluating
develop to become more interpretable, so that Intrusion Detection Systems: The 1998
security teams can understand decision - DARPA Off-Line Intrusion Detection
making, reduce prejudice and improve Evaluation."
reliability [11]. [13] Kruegel, C., et al. (2003). "Anomaly
Detection in Network Intrusion Detection
In addition,A Study on Impacts of Artificial Systems."
Intelligence in Cyber-security side effects pose [14] Chen, T., et al. (2017). "Adversarial
a serious risk and emphasize the need for Machine Learning in Network Intrusion
strong motors. While AI provides sufficient Detection: A Survey."
benefits, human competence remains [15] Wang, H., et al. (2020). "AI and
inevitable, strengthens the importance of a Blockchain: Synergies and Challenges for
hybrid approach that combines AI with human Cyber security."
intelligence for optimal security. Future [16] Cardenas, A. A., et al. (2009). "Challenges
research should focus on improving the for Machine Learning in Anomaly Detection."
flexibility of the AI model, addressing privacy [17] Tran, Q., et al. (2021). "Federated
considerations and reducing calculation costs Learning in Cyber security: A Systematic
to ensure long-term stability in AI operated Review."
cyber security.

8. References
References
[1] Bahnsen, A. C., et al. (2017). "Phishing
Detection with Machine Learning."
[2] Biggio, B., & Roli, F. (2018). "Wild
Patterns: Ten Years after the Rise of
Adversarial Machine Learning."
[3] Buczak, A. L., & Guven, E. (2016). "A
Survey of Data Mining and Machine Learning
Methods for Cyber security."
[4] McAfee Labs (2020). "AI-driven Cyber
security Solutions."
[5] Sharmeen, N., et al. (2021). "AI-based
Incident Response in Cyber security."
[6] Sommer, R., & Paxson, V. (2010). "Outside
the Closed World: On Using Machine Learning
for Network Intrusion Detection."
[7] Good fellow, I., et al. (2015). "Explaining
and Harnessing Adversarial Examples."
[8] Vinaya kumar, R., et al. (2019). "Deep
Learning for Cyber Security Intrusion
Detection: Approaches, Datasets, and
Comparative Study."
[9] Li, Y., et al. (2018). "Cyber Security in the
Era of AI: Risks and Opportunities."
[10] He, Y., et al. (2019). "An Overview of
Machine Learning in Cyber Security."
[11] Saxe, J., & Sanders, K. (2018). "Malware
Data Science: Attack Detection and
Attribution."

IJMSRT25MAY045 www.ijmsrt.com 149

DOI: https://doi.org/10.5281/zenodo.15462132