Abstract

Phishing and social engineering attacks remain major cybersecurity threats by exploiting human behavior rather than system vulnerabilities. This paper investigates artificial intelligence (AI)-driven approaches for detecting such attacks through a combination of literature review, dataset evaluation, system design, and experimental analysis. It examines traditional machine learning models, deep learning techniques, and transformer-based natural language processing methods. A hybrid detection framework is proposed, integrating semantic text analysis, deep learning, and URL-based features within a modular architecture. The system is implemented using a fine-tuned RoBERTa model for email classification and evaluated on multiple benchmark datasets. Results show that transformer-based models achieve high accuracy in identifying contextual and manipulative patterns, while URL-based methods enable fast real-time detection. The findings demonstrate that hybrid approaches significantly enhance detection performance and robustness, particularly against AI-generated phishing attacks, providing a scalable and adaptive solution for modern cybersecurity challenges.

References

• Akeiber, H.J., 2025. The evolution of social engineering attacks: a cybersecurity engineering perspective. Al-Rafidain Journal of Engineering Sciences, pp.294–316.
• Zdrojewski, K., 2025. AI-powered cyberattacks: a comprehensive review and analysis of emerging threats. Advances in IT and Electrical Engineering, 31, pp.55–70.
• Boulieris, P., Pavlopoulos, J., Xenos, A. and Vassalos, V., 2024. Fraud detection with natural language processing. Machine Learning, 113(8), pp.5087–5108.
• Wang, Z., Zhu, H. and Sun, L., 2021. Social engineering in cybersecurity: Effect mechanisms, human vulnerabilities and attack methods. Ieee Access, 9, pp.11895-11910.
• Fakhouri, H.N., Alhadidi, B., Omar, K., Makhadmeh, S.N., Hamad, F. and Halalsheh, N.Z., 2024, February. Ai-driven solutions for social engineering attacks: Detection, prevention, and response. In 2024 2nd international conference on cyber resilience (ICCR) (pp. 1-8). IEEE.
• Tang, L. and Mahmoud, Q.H., 2021. A survey of machine learning-based solutions for phishing website detection. Machine Learning and Knowledge Extraction, 3(3), pp.672-694.
• Alshingiti, Z., Alaqel, R., Al-Muhtadi, J., Haq, Q.E.U., Saleem, K. and Faheem, M.H., 2023. A deep learning-based phishing detection system using CNN, LSTM, and LSTM-CNN. Electronics, 12(1), p.232.
• Jamal, S., Wimmer, H. and Sarker, I.H., 2024. An improved transformer‐based model for detecting phishing, spam and ham emails: A large language model approach. Security and Privacy, 7(5), p.e402.
• Acheampong, F.A., Nunoo-Mensah, H. and Chen, W., 2021. Transformer models for text-based emotion detection: a review of BERT-based approaches. Artificial Intelligence Review, 54(8), pp.5789-5829.
• Veach, A.M. and Abualkibash, M., 2022. Phishing website detection using several machine learning algorithms: a review paper. International Journal of Informatics, Information System and Computer Engineering (INJIISCOM), 3(2), pp.219-230.
• Mirsky, Y. and Lee, W., 2021. The creation and detection of deepfakes: A survey. ACM computing surveys (CSUR), 54(1), pp.1-41.
• Parasaram, V.K.B., 2023. Phishing Detection 2.0: A Natural Language Processing Approach to Identifying Generative AI-Crafted Social Engineering. In International Journal for Research Publication and Seminar (Vol. 14, No. 2, pp. 276-284).
• Devlin, J., Chang, M.W., Lee, K. and Toutanova, K., 2019. BERT: pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pp.4171–4186.
• Sakkis, G., Androutsopoulos, I., Paliouras, G., Karkaletsis, V., Spyropoulos, C.D. and Stamatopoulos, P., 2003. A memory-based approach to anti-spam filtering for mailing lists. Information Retrieval, 6, pp.49–73.
• Klimt, B. and Yang, Y., 2004. The Enron corpus: a new dataset for email classification research. In European Conference on Machine Learning. Berlin/Heidelberg: Springer, pp.217–226.
• Champa, A.I., Rabbi, M.F. and Zibran, M.F., 2024. Curated datasets and feature analysis for phishing email detection with machine learning. In 3rd IEEE International Conference on Computing and Machine Intelligence (ICMI), pp.1–7.
• DeBarr, D. and Wechsler, H., 2012. Spam detection using random boost. Pattern Recognition Letters, 33, pp.1237–1244.
• Nazario, J., 2023. The Online Phishing Corpus. Available at: https://monkey.org/~jose/phishing/ Accessed 20 January 2026.
• Radev, D., 2008. CLAIR Collection of Fraud Email. Available at: https://aclweb.org/aclwiki/CLAIR_collection_of_fraud_email_(Repository) Accessed 20 January 2026.
• Bagui, S., Nandi, D., Bagui, S. and White, R.J., 2021. Machine learning and deep learning for phishing email classification using one-hot encoding. Journal of Computer Science, 17, pp.610–623.
• Shahrivari, V., Darabi, M.M. and Izadi, M., 2020. Phishing detection using machine learning techniques. arXiv, arXiv:2009.11116.
• Salloum, S., Gaber, T., Vadera, S. and Shaalan, K., 2021. Phishing email detection using natural language processing techniques: a literature survey. Procedia Computer Science, 189, pp.19–28.
• Moriya, S. and Shibata, C., 2018. Transfer learning method for very deep CNN for text classification and methods for its evaluation. In Proceedings of the 2018 IEEE 42nd Annual Computer Software and Applications Conference (COMPSAC), IEEE, Tokyo, Japan, 23–27 July 2018, Vol. 2, pp.153–158.
• Khataei Maragheh, H., Gharehchopogh, F.S., Majidzadeh, K. and Sangar, A.B., 2022. A new hybrid based on long short-term memory network with spotted hyena optimization algorithm for multi-label text classification. Mathematics, 10, p.488.
• Eang, C. and Lee, S., 2024. Improving the accuracy and effectiveness of text classification based on the integration of the BERT model and a recurrent neural network (RNN_BERT_Based). Applied Sciences, 14, p.8388.
• Sanh, V., Debut, L., Chaumond, J. and Wolf, T., 2019. DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter. arXiv, arXiv:1910.01108.
• Alhuzali, A., Alloqmani, A., Aljabri, M. and Alharbi, F., 2025. In-depth analysis of phishing email detection: evaluating the performance of machine learning and deep learning models across multiple datasets. Applied Sciences, 15(6), p.3396.