PhishShield: AI-Based Scam Detection and Cyber Awareness System
An intelligent cyber defense and awareness system analyzing messages, links, and screenshots to detect phishing scams with risk scoring and interactive what-if simulations.
Stall Participants & Presenters
Overview
With the rapid proliferation of digital payments, e-commerce, and messaging apps, deceptive phishing vectors, fraudulent SMS lures, and social engineering attacks have become rampant. Everyday users frequently fall victim to spoofed links and fraudulent requests because warning signs are subtle and difficult to decipher without technical training.
PhishShield is an AI-driven scam detection and threat intelligence platform designed to democratize cybersecurity awareness. By combining automated multimodal content analysis with intuitive risk explanations, PhishShield empowers non-technical users to verify suspicious interactions before sharing sensitive credentials or funds, contributing directly to SDG 8 (Decent Work and Economic Growth), SDG 9 (Industry, Innovation, and Infrastructure), and SDG 16 (Peace, Justice, and Strong Institutions).
Multimodal Threat Analysis & Educational Sandboxing
- Multimodal Content Ingestion: Analyzes incoming text messages, suspect URLs, and mobile screenshots to detect fraudulent branding, deceptive typography, and malicious intent.
- Machine Learning Threat Classification: Evaluates submitted artifacts using trained NLP models to classify danger levels into clear strata (Safe, Suspicious, or Dangerous) accompanied by an explainable threat score.
- Deep URL & Domain Inspection: Examines domain age, SSL certificate validity, redirection chains, and lexical homograph spoofing patterns against updated reputation databases.
- Interactive “What-If” Simulator: An educational sandbox demonstrating the operational consequences of clicking a malicious link, submitting an OTP, or authorizing spoofed permissions without exposing user credentials.
- Community Threat Intelligence: Multilingual reporting dashboard allowing users to submit emerging scams, aggregating localized threat patterns for communal defense.
Architecture & Technology Stack
- Frontend & Dashboard: Modern responsive web interface featuring multilingual localization, threat reporting dashboards, and simulator interfaces.
- Machine Learning Core: Supervised NLP classification models and optical character recognition (OCR) pipelines for message text and screenshot analysis.
- URL & Security Scanners: Automated link crawlers and DNS/WHOIS reputation query engines for link structure analysis.
- Backend Services: RESTful API architecture managing ingestion queues, threat scoring calculation, and incident reporting persistence.
- Security & Privacy: Anonymized submission auditing and telemetry protection ensuring user submissions are scrubbed of personal PII.

