In a significant move to bolster user security, Google has introduced advanced artificial intelligence (AI) features designed to detect and prevent scams across its Chrome browser and Android operating system. These enhancements leverage on-device AI models to provide real-time protection against evolving fraudulent activities.
On-Device AI Integration in Chrome
Google’s latest update to Chrome 137 for desktops incorporates Gemini Nano, an on-device large language model (LLM), to enhance the Safe Browsing feature. This integration allows Chrome to instantly analyze web pages for potential threats, offering immediate protection against previously unseen scams. By processing data locally, this approach ensures user privacy while effectively identifying malicious sites.
A primary focus of this enhancement is the detection of remote tech support scams. These scams often deceive users into providing personal or financial information under the guise of resolving non-existent computer issues. The on-device LLM evaluates web pages for indicators typical of such scams, such as the misuse of the keyboard lock API, and communicates these signals to Safe Browsing to assess the likelihood of fraudulent activity.
To maintain optimal browser performance, Google has implemented measures to manage resource consumption. These include limiting the number of tokens processed, running analyses asynchronously to prevent browser interruptions, and enforcing throttling and quotas to control GPU usage.
Looking ahead, Google plans to expand this AI-driven detection to identify other scam types, including those related to package tracking and unpaid tolls. Additionally, the company intends to extend this feature to Chrome on Android devices later this year.
Enhanced Scam Detection in Search Results
Beyond browser security, Google has upgraded its AI-powered systems to identify and filter deceptive web pages from search results more effectively. This advancement has led to a 20-fold increase in the detection of such pages. Notably, in 2024, there was an over 80% reduction in schemes impersonating airline customer service providers and a more than 70% decrease in scams mimicking official resources like visas and government services.
Real-Time Scam Detection in Android Messages
In March 2025, Google introduced AI-powered scam detection features within the Android Messages app. This functionality analyzes message content in real-time, identifying patterns associated with scams and alerting users accordingly. By processing data on-device, this feature ensures user privacy while providing timely warnings about potential threats.
On-Device AI for Scam Call Detection
Google has also rolled out an AI-powered scam detection feature for phone calls on Pixel devices. This system monitors active conversations to detect patterns indicative of scams, such as callers impersonating banks or other institutions requesting urgent fund transfers. When a potential scam is identified, the system provides audio and visual alerts, enabling users to terminate the call promptly. All processing occurs on-device, ensuring that no audio or transcription data is stored or transmitted, thereby maintaining user privacy.
Advanced Theft Protection Features
To further enhance device security, Google has introduced several features aimed at protecting users from theft:
– Theft Detection Lock: Utilizing on-device AI, this feature detects motions associated with theft, such as a sudden snatch, and automatically locks the device screen to prevent unauthorized access.
– Offline Device Lock: If a device remains offline for an extended period, this feature locks the screen to deter thieves from accessing data or circumventing remote wipe capabilities.
– Remote Lock: Users can remotely secure their device by visiting Android.com/lock and entering a verified phone number, even if they cannot access their Google account credentials at the moment.
These theft protection features are available on most Android devices running version 10 or higher and can be enabled through device settings.
Identity Check for Enhanced Security
In January 2025, Google introduced Identity Check, a feature that requires explicit biometric authentication to access sensitive resources when the device is outside of trusted locations. Users can designate trusted locations, and when outside these areas, biometric verification is necessary to change device settings, disable theft protection, or access critical account information. This feature is rolling out to Pixel devices with Android 15 and will be available on other supported Android devices later this year.
Conclusion
Google’s integration of on-device AI across its platforms represents a significant advancement in user security. By processing data locally, these features provide real-time protection against a variety of scams and threats while maintaining user privacy. As fraudulent activities continue to evolve, Google’s proactive approach aims to stay ahead, offering users a safer digital experience.
