Google Deploys On-Device AI to Detect Scam Websites in Chrome

Google is stepping up its fight against online scams with a new artificial intelligence-powered system that can detect fraudulent websites in real time on Chrome. The tech giant announced on Thursday that it’s now using a version of its Gemini AI model directly on users’ devices to identify scammy web pages – particularly those that mimic virus warnings and push fake tech support.

The AI tool is designed to work with Chrome’s enhanced protection mode on desktop, scanning sites as users open them and flagging potential threats before they can do harm. This new system allows Google to identify scam pages that may otherwise evade detection by traditional tools, especially those that use cloaking techniques to hide their true content from Google’s scanners while displaying deceptive material to users.

According to Google, the on-device model, called Gemini Nano, not only speeds up threat detection but also safeguards user privacy by processing data locally instead of sending it to the cloud.

“If a site looks suspicious, we’ll show a warning before you continue,” said Jasika Bawa, group product manager for Chrome. She added that this AI-enhanced defense is more capable of recognizing sophisticated scam tactics and reacting in real time.

The move comes amid growing global concerns about the surge in AI-generated scams. Scammers are now using advanced AI tools to produce convincing fake content, making it easier than ever to deceive people online. In 2023 alone, consumers around the world lost over $1 trillion to various scams, according to the Global Anti-Scam Alliance.

Google has also expanded its AI efforts across other platforms. On Android, users will now get alerts if they receive suspicious website notifications through Chrome and can easily unsubscribe from such sites. In Google Search, AI has significantly boosted scam detection, enabling the company to block 20 times more malicious results than it did just three years ago. In 2024 alone, Google says it removed hundreds of millions of scam pages from search results daily.

One particularly effective use of AI has been in curbing fake customer service pages – especially those targeting airline passengers. The company reported an 80% drop in scams related to airline search results due to improved AI detection.

Google’s senior director of engineering for Search, Phiroze Parakh, acknowledged that the battle against scammers is a constantly evolving one. “Now both sides have powerful tools,” he told CNN. “The challenge is: who can stay a step ahead?”

Google isn’t alone in leveraging AI to fight digital fraud. UK telecom provider O2 has introduced an AI chatbot, Daisy, designed to waste scammers’ time by keeping them engaged in fake conversations. Microsoft has been testing AI systems that analyze phone calls in real time to detect potential scams, while the U.S. Treasury credited AI with helping recover $1 billion in check fraud during the last fiscal year.

As scam tactics evolve, companies like Google are racing to harness the same AI capabilities being used by bad actors – only this time, to protect users instead of deceive them.