Fraud protection: Google integrates local AI model Gemini Nano into Chrome

Google is expanding fraud protection in the Chrome browser. A local AI model is now being used on Android, desktop and search: Gemini Nano.

As the company writes on a blog post, the local AI model Gemini Nano in the desktop browser is designed to protect users from online fraud. Google is also introducing new AI-based warnings for Chrome on Android to warn against spam notifications.

According to Google, the advanced protection mode of the Safe Browsing setting in the Chrome browser represents the "highest level of protection" and protects users "twice as well against phishing and other scams as the standard protection mode". Until now, Google has relied on a server-side URL scan. The on-device approach is intended to provide faster insight into potentially risky websites.

Chrome for Android with protection against spammy notifications

Google also wants to take further measures to mitigate spam notifications. According to Google, if users have enabled notifications from websites, fraudulent sites could try to scam them with a flood of notifications. To protect against malicious, spammy or misleading notifications, Google has introduced AI-powered alerts for Chrome on Android.

As soon as Chrome's AI model detects suspicious notifications on the device, the user will receive a warning with the option to either turn off the notifications or view the blocked content. If the warning was displayed incorrectly, users can decide whether they want to allow future notifications from the website.

Protection against scams in Google search

As part of the announcement, Google also announced that AI would also be used in search to prevent fraud. For example, "hundreds of millions of fraudulent results are detected and blocked every day". According to Google, the AI-supported fraud detection systems have helped to detect 20 times more fraudulent sites.

Among other things, Google had noticed an increase in scammers posing as airline customer service representatives to defraud people seeking help. Google has been able to reduce these scams by over 80 percent, reducing the risk of users coming across a fraudulent phone number when searching.

(afl)