Hackathon Submission | Track: AI for Social Good / Cybersecurity
Seniors lose life savings to Coercion Scams (Digital Arrest). Fraudsters keep victims on a phone call for hours, forcing them to transfer money via apps like GPay or NetBanking. Traditional spam blockers (Truecaller) fail here because the content of the call isn't spam—it's social engineering.
SilverGuard is different. We built a system that detects behavior, not just numbers.
The Coercion Interlock Logic:
IF (Phone Call is ACTIVE) AND (Banking App is OPENED)
THEN -> 🚨 TRIGGER CRITICAL ALERT
This specific pattern—using a payment app while on a call—is 99% indicative of a fraud in progress for elderly users.
- 📱 Real-Time Text Analysis: Reads screen context (WhatsApp/SMS) for risky keywords like "OTP", "Police", "RBI", "Urgent".
- 🔒 Coercion Interlock: Physically detects simultaneous Call + Finance App usage.
- 🛑 Blocking Overlay: A full-screen warning that interrupts the scam flow.
- 👨👩👧 Family Guardian: Instantly sends a Gmail Alert to a family member via N8N cloud webhook.
- 🧠 AI Powered: Uses local scoring + Groq LLaMA-3 for advanced text verification.
- Mobile App: Flutter (Dart)
- Native Modules: Kotlin (AccessibilityService, TelephonyManager)
- AI Backend: Groq API (LLaMA-3-70b)
- Automation: N8N (Webhook -> Gmail/Sheets)
- Install APK: Grant Accessibility and Overlay permissions.
- Test Text Detection: Open WhatsApp and type "This is urgent OTP". The overlay will trigger.
- Test Coercion Interlock:
- Make a phone call to the device.
- While on the call, open Google Pay.
- Watch the CRITICAL RED ALERT trigger immediately.
Built with ❤️ for a Safer Digital India.