Google is significantly ramping up its fraud protection capabilities by integrating advanced AI tools across Search, Chrome, and Android. The updates, powered by its Gemini AI model, are designed to block fake notifications, scam sites, and tech support fraud more effectively and in real time.
Jasika Bawa, Group Product Manager for Chrome, said, “We’re using on-device AI to catch scams faster, smarter, and closer to the user – before any damage is done.” In Google Search, AI now blocks hundreds of millions of scam-related pages each day. According to Bawa, this new system is 20 times more effective at detecting fraudulent content compared to previous tools. One key example is the reduction of fake airline support scams, which have dropped by over 80% due to AI’s ability to detect and suppress fake customer service listings.
Chrome’s Safe Browsing feature now benefits from Gemini Nano, an on-device AI model that helps detect dangerous sites instantly, even those not yet reported. This update is especially effective at stopping tech support scams, which often involve deceptive pop-ups and full-screen warnings. Phiroze Parakh, Senior Director of Engineering, said, “By analyzing pages locally, Chrome can identify scam behavior in real-time without compromising speed or privacy.”
Key features include:
- Detection of hidden or abusive web behaviors
- Real-time threat analysis
- Privacy-preserving on-device scanning
A new layer of protection comes to Chrome for Android, targeting scammy push notifications. These pop-ups often trick users with alarming messages or fake warnings. How does it work? Chrome uses on-device AI to analyze message titles, content, and buttons. Suspicious notifications trigger a warning, offering users the option to unsubscribe. If a site is mistakenly flagged, users can still opt back in. This feature was trained on synthetic data from Gemini LLM and reviewed by human testers, ensuring high accuracy.
In addition to AI features, Chrome includes several built-in tools:
- Abusive Site Detection: Automatically removes notification permissions from bad actors.
- One-Tap Unsubscribe: Lets Android users quickly block unwanted alerts.
- Notification Management: Gives full control over which sites can send push messages.
Chrome version 137 now further enhances protection by analyzing web pages in real-time to detect scam tactics like:
- Locking keyboard input
- Displaying fake security alerts
- Forcing full-screen warnings
Google notes that many scam websites are live for under 10 minutes, often showing different content to real users and security crawlers. Using on-device AI means threats can be detected as users see them, making protections faster and more reliable.
Outside of Chrome, Google is bringing AI-based fraud detection to its Messages and Phone apps. The technology scans texts and call patterns to alert users about potential scams in real time.
Google plans to extend AI protections to address emerging threats, including fake package delivery links, unpaid toll and fine scams, and new web-based fraud tactics. Many of these updates will arrive on Chrome for Android later this year, with broader rollout planned for other platforms.