Google is stepping up its fight against online scams by using artificial intelligence to identify suspicious websites in Chrome. The company announced it’s now leveraging a version of its Gemini AI model, called Gemini Nano, directly on users’ devices to detect and alert users to common scams—like fake virus warnings urging users to click for tech support or download software.
This is part of a broader effort by Google to enhance user protection using AI across Chrome, Search, and Android. The move comes as scammers increasingly exploit AI to create more convincing and widespread fraud attempts. According to the Global Anti-Scam Alliance, victims lost over $1 trillion globally to scams last year.
“Fighting scammers is a constant game of evolution,” said Phiroze Parakh, a senior engineering director at Google Search. As AI tools become more accessible to both scammers and tech companies, the challenge is to stay one step ahead.
While Google has long relied on machine learning for safety, newer AI models with better language understanding and pattern recognition are helping identify scams more quickly and accurately. On Chrome’s “enhanced protection” mode for desktop, Gemini Nano now scans pages in real time after a user clicks on them, offering faster detection and more privacy since it runs on-device.
The AI also combats a tactic called “cloaking,” where scammers disguise websites to fool Google’s crawlers while showing something different to users. Chrome now warns users before they enter these suspicious sites.
In addition, Android users who receive unwanted site notifications in Chrome will now get alerts and can automatically unsubscribe—if Chrome notifications are enabled.
Beyond Chrome, Google has strengthened its AI-powered systems in Search to block scammy results. The company says it’s now blocking 20 times more scam pages than before, including a major reduction in fake airline customer service pages—cutting such scams by 80%.
Google isn’t alone in this effort. UK telecom O2 uses an AI bot named “Daisy” to keep phone scammers talking and waste their time. Microsoft is testing an AI tool to flag potential phone scams in real time, and the U.S. Treasury credits AI for recovering $1 billion in check fraud during fiscal year 2024.
