Evaluating the Behavioural Impact of Machine Learning-Driven Cybersecurity Awareness Programs in Telecommunication Networks

Abstract

The increasing frequency and sophistication of cyberattacks on telecommunication networks require sophisticated techniques for raising the awareness of users and encouraging secure online conduct. This study explores the behavioural effects of machine learning-based cybersecurity awareness programs deployed on telecommunication and computer systems. The main objective is to assess the impact of these data-informed programs on participants' awareness of cyber threats and encourage the uptake of secure behaviours, such as the use of strong passwords, detection of phishing attacks, and secure management of personal information. A mixed-methods design was used, combining pre- and post-program surveys, statistical analysis using SPSS (paired t-tests, chi-square), and transformer-based analysis with Explainable AI (XAI). The research included 500 participants across five telecommunication networks, chosen using stratified random sampling to provide representative findings in technical and non-technical areas. Statistically significant increases in cybersecurity knowledge (t = 8.76, p < 0.001) and behaviour were found, with significant increases in high-awareness scores (from 25% to 60%) and in important actions like turning on multi-factor authentication and not clicking on phishing links. BERT-XAI analysis indicated a 63% increase in open-ended mentions of proactive security practices, providing interpretable, individualized behavioural insights. The results show the promise of machine learning-based awareness programs to deliver adaptive and quantifiable effects on user behaviour. The findings affirm the embedding of smart education models within telecommunication infrastructures and offer real-world advice for scalable, data-driven cyber education strategy design that continues to prove effective and valid in the face of changing digital threats.