Google tests Gemini antiscam protection in Chrome

Google is testing Antiscam protection using Gemini in Chrome, with the feature tied to Enhanced Safe Browsing.

Chrome already protects users from phishing and tech support scams using AI-based detection when Enhanced Safe Browsing is enabled. These protections rely on a mix of local checks and server-side signals to identify suspicious pages before they can cause harm.

Gemini Antiscam Protection in Chrome

Google is extending those protections by using Gemini as a server-side system to assess whether a web page is linked to scam activity.

The system activates only when Chrome detects a page that already raises suspicion. Rather than scanning every site a user visits, Chrome steps in at higher-risk moments and sends the page for deeper analysis. This keeps the process targeted and adds another layer to Chrome’s existing scam checks without replacing them.

Under the hood, Chrome uses a server-side Gemini model to evaluate how likely a page is to be a scam. The analysis includes the page’s URL and visible text content and is sent to a Google-owned service for processing. Cookies are not allowed, and the feature operates only when Enhanced protection is turned on in Safe Browsing. The feature is disabled by default and requires Enhanced Safe Browsing to be enabled.

As spotted by Windows Report, Chrome includes an experimental flag in Canary called Gemini Antiscam Protection. The commit links it to Enhanced Safe Browsing and points to a possible release timeframe around Chrome 146 across Windows, macOS, Linux, ChromeOS, and Android.

Chrome makes it easier to ask AI about webpages, tests a startup option on Windows, and plans an AI disclosure label for websites.