Google Rolls Out On-Device AI Protections to Detect Scams in Chrome and Android
Google on Thursday announced it’s rolling out new artificial intelligence (AI)-powered countermeasures to combat scams across Chrome, Search, and Android.
The tech giant said it will begin using Gemini Nano, its on-device large language model (LLM), to improve Safe Browsing in Chrome 137 on desktops.
“The on-device approach provides instant insight on risky websites and allows us to offer protection, even against scams that haven’t been seen before. Gemini Nano’s LLM is perfect for this use because of its ability to distill the varied, complex nature of websites, helping us adapt to new scam tactics more quickly,” the company said.
Google noted that it’s already using this AI-driven approach to tackle remote tech support scams, which often seek to trick users into parting with their personal or financial information under the pretext of a non-existent computer problem.

This works by evaluating the web pages using the LLM for potential signals that are emblematic of tech support scams, such as the use of the keyboard lock API. The security signals are then extracted and passed to Safe Browsing to determine if the page is likely a scam.
“In addition to ensuring that the LLM is only triggered sparingly and run locally on the device, we carefully manage resource consumption by considering the number of tokens used, running the process asynchronously to avoid interrupting browser activity, and implementing throttling and quota enforcement mechanisms to limit GPU usage,” Jasika Bawa, Andy Lim, and Xinghui Lu of the Google Chrome Security team said.
Google said it intends to expand this feature to detect other kinds of scams, including those related to package tracking and unpaid tolls. The feature is also expected to be rolled out to Chrome on Android later this year.

As part of the announcement, Google also revealed that it has enhanced its AI-powered scam detection systems to ensnare 20 times more deceptive pages and block such pages from search results, reducing schemes that impersonate airline customer service providers by over 80% and those that mimic official resources like visas and government services by over 70% in 2024.
Lastly, Google said it’s launching a new warnings feature for Chrome on Android that uses an on-device machine learning model to alert users of unwanted notifications sent by malicious sites that aim to trick them into downloading suspicious software or providing sensitive data.
“This new feature uses on-device machine learning to detect and warn you about potentially deceptive or spammy notifications, giving you an extra level of control over the information displayed on your device,” Chrome Security’s Hannah Buonomo and Sarah Krakowiak Criel said.
“When a notification is flagged by Chrome, you’ll see the name of the site sending the notification, a message warning that the contents of the notification are potentially deceptive or spammy, and the option to either unsubscribe from the site or see the flagged content.”

The features come a little over two months after Google rolled out AI-powered scam detection features in the Messages app for Android. Last year, the company unveiled similar ways to flag scam calls.
The updates also arrive as Google appears to be readying an Advanced Protection feature in Android 16 that, in some ways, mirrors Apple’s approach by turning off JavaScript, disabling 2G connections, and activating a number of security features by default, such as Theft Detection Lock, Offline Device Lock, Android Safe Browsing, spam protection in Messages.
Google has also been spotted working on a feature to detect scams that coax victims into opening their banking apps during phone calls, Android Authority reported earlier this week.