Google is using AI to identify scammy websites on Chrome when you click on them

Almost anyone who has used the internet has encountered a moment where a pop-up claims their device has a virus, urging them to click for support or download suspicious software. These fake alerts are part of a widespread tech support scam that Google is now tackling more aggressively with artificial intelligence. The company is using a version of its Gemini AI model, called Gemini Nano, which runs directly on users’ devices to detect and warn against scam websites in real time.

Running the AI on-device helps Google provide faster protection while preserving user privacy. This feature is integrated into Chrome’s Enhanced Protection safe browsing mode. When users click on a potentially harmful link, the AI scans the page in real time and displays a warning if it detects deceptive content. This is especially helpful in identifying tactics like “cloaking,” where scam sites present different content to Google’s web crawlers than they do to real users.

Beyond Chrome, Google is also using AI to improve security in its Search and Android platforms. In Search, new AI systems have helped block twenty times more scammy web pages than before. For example, scams involving fake airline customer service pages have dropped by 80% due to better scam detection. On Android, users will soon receive alerts when scammy websites try to send notifications through Chrome, with the option to unsubscribe instantly.

Google says these improvements are part of a broader effort to combat increasingly sophisticated scams made possible by advances in generative AI. Scammers are now able to create large volumes of convincing fake content quickly and cheaply, making traditional defense mechanisms less effective. But Google’s new AI models, which offer better language understanding and context recognition, are helping level the playing field.

Other organizations are taking similar steps. UK telecom company O2 is using an AI chatbot to tie up scammers on calls, and Microsoft has introduced tools that analyze phone conversations to flag potential fraud. In the United States, the Treasury Department reported using AI to help recover $1 billion in check fraud during fiscal year 2024.

As scammers become more advanced, tech companies are racing to develop tools that can keep pace. Google’s use of on-device AI in Chrome, Search, and Android reflects a growing reliance on artificial intelligence not just to power user experiences, but to protect them.