Google Chrome uses AI to analyze pages in new scam detection feature

Google is using artificial intelligence to power a new Chrome scam protection feature that analyzes brands and the intent of pages as you browse the web.

As spotted by Leo on X, a new flag in Chrome Canary enables a feature called "Client Side Detection Brand and Intent for Scam Detection" that uses an LLM, or Large Language Model. to analyze web pages on your device.

"Enables on device LLM output on pages to inquire for brand and intent of the page," reads the Google Chrome flag's description.

This feature is believed to help the scam detection service detect the brand and purpose (intent) of a webpage, making it easier to identify potential scams. It works on Mac, Windows, and Linux.

It's unclear how the feature works, but it could issue warnings when you visit an obvious scam website.

For example, if you visit a fake Microsoft tech support page claiming your computer is infected and urging you to call a number, Chrome's AI could analyze the promoted brand or language used on the page. If it detects scam tactics like fake urgency or suspicious domains, it could display a warning alerting you to avoid interacting with the page or sharing personal information.

This new tool is being tested in Chrome Canary and could be related to Chrome's built-in Enhanced Protection feature, which now also uses artificial intelligence.

Chrome's Enhanced Protection is now powered by AI.

Google says the updated Enhanced Protection feature uses AI to provide real-time protection against dangerous sites, downloads, and extensions.

Before October, Enhanced Protection didn't use AI. It was described as "proactive protection," but it has since been updated to "AI-powered protection."

Google is likely using pre-trained data to understand web content and warn users about scams or dangerous sites.

The company is still testing these AI-powered security and privacy features in Chrome, and it's unclear when more details will be shared.