Google’s AI-Driven Safety Tools for Scam Detection: How it Works

AI-Powered Real-Time Scam Detection: Google embeds the Gemini Nano on-device AI model into Pixel devices (starting with Pixel 9), where it analyses incoming calls in real time. The system listens for conversational patterns of common scams and flags suspicious calls without ever sending audio to Google.

Privacy-First Design: All call analysis happens locally on the device. No call is recorded, transcribed, or uploaded to Google’s servers. This ensures that your voice remains private, while still benefiting from powerful scam protection.

Real-Time Audible & Visual Alerts: When the AI detects scam-like behavior, it triggers an audible beep during the call and shows a warning on-screen. This alert gives you the option to disconnect immediately if something feels off.

User Control & Opt-In Feature: Scam detection is off by default. Users can turn it on manually, and it only works for calls from unsaved (unknown) numbers, so trusted contacts aren’t wrongly flagged.

Protection During Screen Sharing: If you share your screen while on a call with an unknown person, and you open sensitive financial apps like Google Pay, Paytm, or Navi, Google warns you with a banner. You can end the call and stop screen sharing with a single tap, reducing real-time fraud risk.

Broader AI Safety Ecosystem: Google’s scam protection goes beyond calls: it also uses Gemini Nano in Chrome and Search to detect phishing, spammy sites, and scammy notifications.

Strategic & Layered Fraud Prevention: Google complements AI call detection with other safety measures: SIM-based phone verification (ePNV) to replace SMS OTPs, partnerships with fintech firms, and user education to boost fraud awareness, especially in high-risk markets like India.

Join our WhatsApp Channel to get the latest news, exclusives and videos on WhatsApp