AI-Assisted Cybercrime Management System for Somalia

Abstract

The growth of digital services has increased both the scale and complexity of cybercrime, while cybercrime handling in Somalia remains constrained by fragmented reporting and largely manual investigative processes. This paper presents the design, implementation, and evaluation of an AI-powered cybercrime tracking system for centralized reporting, case management, digital-evidence handling, and investigation support in Mogadishu, Somalia. The system was implemented as a full-stack web application using HTML, CSS, and JavaScript for the frontend; Node.js and Express.js for the backend; MongoDB for data storage; and the OpenAI API for AI-assisted case categorization, suspect risk scoring, evidence analysis, victim-statement analysis, similar-case detection, and report generation. The platform provides role-based portals for citizens, analysts, investigators, and administrators, together with JWT-based authentication, SHA-256 evidence integrity verification, IP geolocation records, automated notifications, and PDF case reports. Requirements were informed through observation and interviews, while feasibility was assessed across technical, operational, economic, and schedule dimensions. Functional testing with Postman, verification of evidence hashing and tamper detection, AI-summary testing, and cloud deployment demonstrated the operational feasibility of the implemented platform. The reported evaluation indicates improvements in cybercrime documentation, case organization, and investigation support compared with manual record-keeping. However, quantitative measures of categorization accuracy, response time, false-positive rate, and user acceptance were not collected in this study and are identified as priorities for future evaluation. The study provides a practical, low-cost architecture for AI-assisted cybercrime management in a resource-constrained context and identifies directions for longitudinal evaluation, multi-institutional deployment, and more advanced predictive analytics.