AN AI DRIVEN MULTIMEDIA FAULT REPORTING AND ANTI-VANDALISM SYSTEM FOR MALAWI’S PUBLIC INFRASTRUCTURES

Public infrastructure faults and vandalism can disrupt services, increase repair costs, and delay restoration. In Malawi, incident reporting may depend on calls, written notices, and manual coordination, hindering consistent evidence collection and prompt location sharing. This paper presents an AI-Driven Multimedia Fault Reporting and Anti-Vandalism System designed as a software platform for citizens, administrators, and public institutions. Accessible through a web platform and mobile application, it allows users to submit incident descriptions, photographs, audio files, and location information. AI-assisted classification analyzes submitted reports to help distinguish technical faults from suspected vandalism and supports severity assessment and administrative review. The platform includes offline report preparation and synchronization, geolocation mapping, report tracking, notifications, and an Emergency Response Module to prioritize critical reports.

The system uses a centralized workflow in which reports are stored, reviewed, classified, and made available to authorized administrators. Role-based access control limits access to functions and records according to user roles. Multimedia uploads are linked to incident records, while mapping tools allow users to provide device-based location information when permitted or select a location manually. Offline reporting supports preparation without connectivity, with pending information synchronized when access returns. These functions aim to improve report completeness and make incident progress easier to monitor.

The system uses existing user devices and ordinary connectivity; it requires no dedicated cameras, sensors, or Internet of Things installations. Development followed an Agile approach. The project documents functional, security, performance, recovery, and usability testing. The test plan marks functional, security, performance, and recovery testing as tested, while usability testing is recorded as not fully tested. The report provides no measured detection-accuracy or response-time results, so effectiveness requires further evaluation. Overall, the system offers a workflow intended to improve report quality, strengthen coordination, support accountable follow-up, and help public institutions respond to infrastructure faults and suspected vandalism.