Google announced on Thursday it is introducing advanced AI-powered protections designed to combat scams for Chrome users. Leveraging the company’s Gemini Nano, an on-device large language model (LLM), these enhanced defenses will initially roll out to desktop users, providing a stronger shield against online scams.
The company highlighted that Chrome’s Enhanced Protection mode already offers double the security against phishing and similar threats compared to its Standard counterpart. Now, by integrating Gemini Nano, Google is enhancing this layer further, equipping users to block even novel scams that haven’t been previously seen.
In a post detailing the announcement, Google emphasized Gemini Nano’s suitability for the task, particularly due to its capability to swiftly analyze and adapt to new and increasingly complex scamming methods. Currently, this system is actively safeguarding Chrome users from fraudulent remote tech support websites, and Google plans to soon extend protection to Android devices and additional scam types.
Alongside this desktop initiative, Google is rolling out new AI-driven warnings for Chrome notifications on Android. Malicious websites sometimes exploit push notifications, beyond their own pages, to target users with scams or misleading content. When Chrome’s AI model identifies notifications as suspicious, users will now receive explicit alerts, giving them the option either to block further notifications or to proceed despite the risk. Users can also override these warnings if needed.
Google also revealed broader applications of AI technology within its platform, saying it has ramped up using AI to purge hundreds of millions of scam-ridden search results daily. The company’s enhanced AI defenses have reduced scam results in Google Search by a factor of twenty, significantly improving online security for users.
One example highlighted was Google’s successful crackdown on scams involving fake airline customer service numbers listed in search results. Thanks to these new AI tactics, the prevalence of such fraudulent listings has fallen by over 80%, greatly minimizing risks for people seeking genuine customer support online.
