BulliShield: A Smart Cyberbullying Detection and Reporting System
Farhan Ishrak Tahmid, Farhana Akbar, Ahsanur Rahman · 2024
BulliShield is an intelligent and portable application that combines detection, reporting, and AI-based counseling to offer an integrated approach to solve the widespread problem of cyberbullying. Within the system's user-friendly architecture, registered users can file complaints about possible instances of cyberbullying. Since institutions typically require the victim to submit photo evidence showing a snapshot of the whole electronic conversation with the bully, BulliShield allows victims to submit photo evidence of conversations with the bully. Our app extracts texts from those photos using OCR, identifies the language (currently, it can only detect hate speech in Bangla and English texts), and then employs appropriate machine/deep learning techniques to identify potentially offensive content in those texts, thereby helping the appropriate authority (for example, a proctor in the case of a university). BulliShield allows authorities to review complaints, call both parties for meetings, add notes, save their judgments to resolve the case, and so on – thereby acting as a fully functional system for handling cyberbullying cases. BulliShield ensures accountability and transparency throughout the resolution process by keeping the plaintiff informed about the current status of the complaint. BulliShield also offers an AI-based counseling component, giving the victims easy and free access to a technology that can help them develop resilience and emotional well-being.