Open Access

ARTICLE

AI-Driven Cybersecurity Framework for Safeguarding University Networks from Emerging Threats

Boniface Wambui^1,*, Margaret Mwinji¹, Hellen Nyambura²

1 School of Computing and Informatics, Mount Kenya University, Thika, 342-01000, Kenya
2 School of ICT and Engineering, Zetech University, Nairobi, 2768-00200, Kenya

* Corresponding Author: Boniface Wambui. Email:

Journal of Cyber Security 2025, 7, 463-482. https://doi.org/10.32604/jcs.2025.069444

Received 23 June 2025; Accepted 24 September 2025; Issue published 23 October 2025

Abstract

As universities rapidly embrace digital transformation, their growing dependence on interconnected systems for academic, research, and administrative operations has significantly heightened their exposure to sophisticated cyber threats. Traditional defenses such as firewalls and signature-based intrusion detection systems have proven inadequate against evolving attacks like malware, phishing, ransomware, and advanced persistent threats (APTs). This growing complexity necessitates intelligent, adaptive, and anticipatory cybersecurity strategies. Artificial Intelligence (AI) offers a transformative approach by enabling automated threat detection, anomaly identification, and real-time incident response. This study sought to design and evaluate an AI-driven cybersecurity framework specifically for university networks in Kenya, focusing on detecting, preventing, and mitigating emerging cyber risks. Utilizing machine learning and deep learning techniques, the framework analyzed extensive network traffic to uncover anomalies and predict potential breaches. Data was collected from 150 university staff members, yielding 120 valid responses, with faculty and students being key participants. Findings showed high awareness (84.2%) and concern (74.2%) about cybersecurity, with phishing (38.3%) and unauthorized access (20%) reported as the most frequent threats. Although 96.7% of respondents employed strong passwords and multi-factor authentication (MFA), only 65.8% considered institutional cybersecurity training adequate. Notably, 95.9% supported AI-driven real-time threat detection, and 93.3% trusted AI to reduce unauthorized access. Statistical analysis revealed a moderate positive correlation (R = 0.466) between cybersecurity awareness and perceived AI effectiveness. The study emphasizes the urgent need for universities to integrate AI-powered security systems with continuous training programs to enhance resilience and create a safer academic digital environment.

---

Keywords

---