A Machine Learning Approach to Real-Time Email Security: Design and Implementation of Browser Extension for Threat Analysis and Modality Scoring
Student: Mohammed-mustapha Adebayo Yusuf (Project, 2025)
Department of Information Technology
Bayero University, Kano, Kano State
Abstract
Email communication is a vital tool for personal and professional interactions, yet its widespread use has made it a prime target for cyberattacks, including phishing, misuse of disposable emails, malware propagation, and social engineering. Spammers often exploit email systems to distribute questionable content, necessitating advanced detection mechanisms to safeguard sensitive information. Traditional email security systems, such as spam filters, frequently fail to address sophisticated and evolving threats, particularly in regions with low cyber literacy, user negligence, and limited technological advancements. This study proposes the development of a browser extension leveraging machine learning to enhance real-time email security. The project aims to design and implement a system capable of detecting malicious emails by identifying disposable email addresses as a frontline security measure, analyzing content, links, and attachments, all while maintaining user privacy through non-persistent data analysis. The methodology integrates a machine learning model into a cross-browser extension compatible with Chrome, Firefox, and Edge, utilizing the WebExtensions API for seamless performance. The system evaluates emails in real time, classifies threats using a modality scoring system, and delivers severity alerts to users. Key features include Natural Language Processing (NLP) for content analysis, link and attachment scanning, and an intuitive interface for threat visualization. Results demonstrate the extension’s effectiveness in identifying malicious content with minimal false positives, making it a reliable tool for under-resourced and high-risk regions. The study concludes that AI-driven solutions embedded in browser extensions offer scalable and cost-effective approaches to enhancing email security. Recommendations include refining the machine learning model for improved accuracy, expanding compatibility to additional browsers for broader accessibility, and implementing cybersecurity awareness campaigns to complement technical interventions.
Keywords
For the full publication, please contact the author directly at: yusufmustapha349@gmail.com
Filters
Institutions
- Federal College of Education (Tech), Gusau, (Affl To Abu Zaria), Zamfara State 1
- Federal College of Education, Abeokuta (Aff To University of Ibadan), Ogun State 2
- Federal College of Education, Eha-Amufu, Enugu State 1
- Federal College of Education, Kano (Affl To Ahmadu Bello University, Zaria) 1
- Federal College of Education, Kontagora, (Affl To Abu, Zaria), Niger State 2
- Federal College of Education, Okene, (Affl. To University of Ibadan), Kogi State 3
- Federal College of Education, Pankshin, (Affl To Uni of Jos), Plateau State 2
- Federal College of Education, Zaria, Kaduna State (affl To Abu, Zaria) 1
- Federal College of Wildlife Management, New Bussa, Niger State 1
- Federal Cooperative College, Ibadan P.m.b. 5033, Eleyele, Ibadan, Oyo State 3