Ddos 
Image: Google
In an era where online deception moves at the speed of code, Google is turning to artificial intelligence to counter tech support scams—one of the most deceptive and persistent cyber threats haunting the modern web.
In a new blog post released by Google, the company outlines a transformative step in Chrome 137: the integration of on-device protection powered by the Gemini Nano large language model (LLM) to detect and block tech support scams. These scams, which attempt to trick users into paying for fake services or surrendering control of their devices, have grown increasingly sophisticated.
“In a tech support scam, the goal of the scammer is to trick you into believing your computer has a serious problem… and then convince you to pay for unnecessary services,” Google explains. The ruse often involves alarming pop-ups, full-screen takeovers, and even the disabling of keyboard and mouse input to simulate a crisis and push victims into panic mode.
Until now, Chrome relied heavily on the cloud-powered Safe Browsing service, which blacklists known malicious URLs. But in a threat landscape where the average malicious site exists for less than 10 minutes, that’s often not fast enough. Enter on-device AI.
“On-device protection allows us to detect and block attacks that haven’t been crawled before,” the post emphasizes. “It also empowers us to see threats the way users see them,” accounting for evasive tactics that display different content based on the user’s device or behavior.
When Chrome suspects a page might be dangerous—such as when it detects the use of the keyboard lock API, a hallmark of scam sites—it calls on Gemini Nano. The lightweight LLM analyzes the content locally, extracting key security signals such as malicious intent. These insights are then sent to Safe Browsing.
Worried about privacy or lag? Google says you shouldn’t be. The AI is run entirely locally on the device and triggered sparingly to avoid draining system resources. The LLM operates asynchronously, minimizing interference with browsing, and is throttled to manage GPU usage. Importantly, summarized security signals are only sent to Google if the user has opted into Enhanced Protection mode.
Even those using Standard Protection will benefit: once a scam site is flagged, it’s added to the global Safe Browsing blocklist, helping others avoid the trap.
Related Posts:
- Hackers attack BitGrail Exchange and steal $170 million Nano
- Google Boosts Real-Time Protection Against Scams and Malware on Android Devices
- AI Dev Gallery: Microsoft Unleashes On-Device AI for Windows 11
- Pro vs. Free: Gemini 2.5’s Tiered AI Power
Donate