The rapid adoption of Artificial Intelligence (AI) systems in critical sectors of society has given rise to new cybersecurity challenges. Unlike traditional software systems, AI systems have unique characteristics such as data dependence, model complexity, and adaptive behavior, which create new types of vulnerabilities and attack vectors. Through such attacks, intruders can manipulate these systems to change their behavior to achieve their goals. According to expert data, only 25% of modern artificial intelligence applications are properly protected. Given these technologies specifics their security covers a wide range of tasks, including the data protection, algorithmic models and application scenarios. This review article provides a comprehensive analysis of the current state of AI cybersecurity, systematizing the vulnerabilities inherent in AI, classifying the main types of attacks at different stages of the AI lifecycle, and describing adequate countermeasures. This paper proposes a comprehensive taxonomy of threats and defenses, covering aspects from data collection to model deployment and operation. The goal of the paper is to provide a deep understanding of the complex AI threat landscape and guide researchers and practitioners in the development and implementation of robust and secure AI systems. Finally, current research gaps are identified and future directions are outlined to ensure the sustainability of AI in a dynamically changing digital environment.
References
[1]
Akhter, S., Ahmad, M.R., Chibb, M., Zai, A.F. and Yaqoob, M. (2024) Artificial Intelligence in the 21st Century: Opportunities, Risks and Ethical Imperatives. EducationalAdministration: TheoryandPractice, 30, 4600-4605. https://doi.org/10.53555/kuey.v30i5.3125
[2]
Rashid, A.B. and Kausik, M.A.K. (2024) AI Revolutionizing Industries Worldwide: A Comprehensive Overview of Its Diverse Applications. HybridAdvances, 7, Article ID: 100277. https://doi.org/10.1016/j.hybadv.2024.100277
[3]
Klishin, A.A. and Taran, K.K. (2024) Legal Issues of Cybersecurity, Risks and Ethics in the Use of Artificial Intelligence. Law and Economics, No. 10, 24-30.
[4]
Mahatme, J. and Aher, P.S. (2025) Advancements in Machine Learning: Revolutionizing the Future of Technology. International Journal of Research and Analytical Reviews, 12, 109-114.
[5]
Tuoyo, O.S., Hossain, A., Habibur Rahman, H.B., Al Mamun, M.A., Hussein, L., Khan, M.A., Melon, M.M.H. and Shah, S. (2024) The Role of Machine Learning and Deep Learning in Shaping Modern Computer Science: Challenge, Opportunities, and Future Directions. Nanotechnology Perceptions, 20, 219-235.
[6]
Sangwan, R.S., Badr, Y. and Srinivasan, S.M. (2023) Cybersecurity for AI Systems: A Survey. JournalofCybersecurityandPrivacy, 3, 166-190. https://doi.org/10.3390/jcp3020010
[7]
Elahi, M., Afolaranmi, S.O., Martinez Lastra, J.L. and Perez Garcia, J.A. (2023) A Comprehensive Literature Review of the Applications of AI Techniques through the Lifecycle of Industrial Equipment. DiscoverArtificialIntelligence, 3, Article No. 43. https://doi.org/10.1007/s44163-023-00089-x
[8]
Hassija, V., Chamola, V., Mahapatra, A., Singal, A., Goel, D., Huang, K., et al. (2023) Interpreting Black-Box Models: A Review on Explainable Artificial Intelligence. CognitiveComputation, 16, 45-74. https://doi.org/10.1007/s12559-023-10179-8
[9]
Christoper Osazuwa, O. and Ozohu Musa, M. (2024) The Expanding Attack Surface: Securing AI and Machine Learning Systems in Security Operations. InternationalJournalofInnovativeScienceandResearchTechnology (IJISRT), 9, 2498-2505. https://doi.org/10.38124/ijisrt/ijisrt24may1613
[10]
Hamon, R., Junklewitz, H., Soler Garrido, J. and Sanchez, I. (2024) Three Challenges to Secure AI Systems in the Context of AI Regulations. IEEEAccess, 12, 61022-61035. https://doi.org/10.1109/access.2024.3391021
[11]
CROWDSTRIKE 2025 Global Threat Report. https://go.crowdstrike.com/2025-global-threat-report.html
[12]
Dhanaraj, A. (2025) The Evolution of Cyber Threats: From Traditional Attacks to Ai-Powered Challenges. EuropeanJournalofComputerScienceandInformationTechnology, 13, 50-61. https://doi.org/10.37745/ejcsit.2013/vol13n365061
[13]
Scott, J. and Kyobe, M. (2021) Trends in Cybersecurity Management Issues Related to Human Behaviour and Machine Learning. 2021 International Conference on Electrical, Computer and Energy Technologies (ICECET), Cape Town, 9-10 December 2021, 1-8. https://doi.org/10.1109/icecet52533.2021.9698626
[14]
Wiafe, I., Koranteng, F.N., Obeng, E.N., Assyne, N., Wiafe, A. and Gulliver, S.R. (2020) Artificial Intelligence for Cybersecurity: A Systematic Mapping of Literature. IEEEAccess, 8, 146598-146612. https://doi.org/10.1109/access.2020.3013145
[15]
Huang, M. and Rust, R.T. (2018) Artificial Intelligence in Service. JournalofServiceResearch, 21, 155-172. https://doi.org/10.1177/1094670517752459
[16]
Joseph, A.D., Laskov, P., Roli, F., Tygar, J.D. and Nelson, B. (2013) Machine Learning Methods for Computer Security (Dagstuhl Perspectives Workshop 12371). Dagstuhl Manifestos, 3, 1-30.
[17]
Qiu, J., Wu, Q., Ding, G., Xu, Y. and Feng, S. (2016) A Survey of Machine Learning for Big Data Processing. EURASIPJournalonAdvancesinSignalProcessing, 2016, Article No. 67. https://doi.org/10.1186/s13634-016-0355-x
[18]
Kaplan, A. and Haenlein, M. (2019) Siri, Siri, in My Hand: Who’s the Fairest in the Land? On the Interpretations, Illustrations, and Implications of Artificial Intelligence. BusinessHorizons, 62, 15-25. https://doi.org/10.1016/j.bushor.2018.08.004
[19]
Abbas, N.N., Ahmed, T., Shah, S.H.U., Omar, M. and Park, H.W. (2019) Investigating the Applications of Artificial Intelligence in Cyber Security. Scientometrics, 121, 1189-1211. https://doi.org/10.1007/s11192-019-03222-9
[20]
Zhang, Z., Ning, H., Shi, F., Farha, F., Xu, Y., Xu, J., Choo, K.K.R., et al. (2021) Artificial Intelligence in Cyber Security: Research Advances, Challenges, and Opportunities. Artificial Intelligence Review, 55, 1029-1053.
[21]
Kumar, P., Wazid, M., Singh, D.P., Singh, J., Das, A.K., Park, Y., et al. (2023) Explainable Artificial Intelligence Envisioned Security Mechanism for Cyber Threat Hunting. SecurityandPrivacy, 6, 112-119. https://doi.org/10.1002/spy2.312
[22]
Leonteva, L. (2025) Evaluating Adversarial Attacks against Artificial Intelligence Systems in Application Deployments. AppliedAILetters, 6, 45-52. https://doi.org/10.1002/ail2.121
[23]
Namiot, D.E. and Ilyushin, E.A. (2022) On the Stability and Security of Artificial Intelligence Systems. International Journal of Open Information Technologies, 10, 126-134.
[24]
Li, Y. and Liu, Q. (2021) A Comprehensive Review Study of Cyber-Attacks and Cyber Security; Emerging Trends and Recent Developments. EnergyReports, 7, 8176-8186. https://doi.org/10.1016/j.egyr.2021.08.126
[25]
Yang, C., Yang, Y. and Zhang, Y. (2024) Understanding the Impact of Artificial Intelligence on the Justice of Charitable Giving: The Moderating Role of Trust and Regulatory Orientation. JournalofConsumerBehaviour, 23, 2624-2636. https://doi.org/10.1002/cb.2365
[26]
Garbuk, S.V. (2024) Special Security Model for the Creation and Use of Artificial Intelligence Systems. Cybersecurity Issues, No. 1, 15-23.
[27]
Yang, J., Zheng, J., Zhang, Z., Chen, Q.I., Wong, D.S. and Li, Y. (2022) Security of Federated Learning for Cloud-Edge Intelligence Collaborative Computing. InternationalJournalofIntelligentSystems, 37, 9290-9308. https://doi.org/10.1002/int.22992
[28]
Nair, P. and Ansari, M.F. (2024) Vulnerabilities in AI Systems: The Integration of AI into Cybersecurity Tools and Systems. International Research Journal of Engineering and Technology (IRJET), 11, 1159-1160.
[29]
Agha, A. (2024) Artificial Intelligence for Vulnerability Management: Enhancing Security in a Dynamic Threat Landscape. TIJER-International Research Journal, 11, a430-a432.
