Google announced on Thursday that it would be releasing new artificial intelligence ( AI)-based measures to stop schemes in Google’s Chrome, Search, and Android.
The tech giant announced that it will begin using , its large language model ( LLM) for Chrome 137 desktops, to improve .
” The on-device method allows us to give protection, even against scams that haven’t already been discovered,” says the company. The organization Gemini Nano’s LLM is ideal for this purpose because it enables us to quickly adapt to new scam strategies because it can extract the complex, intricate nature of websites.
Google noted that it is already employing this AI-driven strategy to combat remote tech support schemes, which frequently deceive customers into giving up their personal or financial details under the pretext of a non-existent computer issue.
This works by using the LLM to look for possible indicators of tech support schemes, such as the use of the console switch API, when the web pages are being examined for these indicators. Safe Browsing uses the safety signals to determine whether the website is a scam after which it is extracted and passed on.
The Google Chrome Security team’s Jasika Bawa, Andy Lim, and Xinghui Lu of the company ,” We carefully control tool usage by considering the number of tokens used, running the process simultaneously to avoid disrupting browser activity, and implementing throttling and quota enforcement mechanisms to reduce GPU usage.
Google said it plans to add additional categories of schemes, including those involving package tracking and underpaid tolls. After this year, the function is anticipated to be made available for Chrome on Android.
Google also made the announcement, revealing that it has improved its AI-powered fraud monitoring systems to entangle 20 times more dishonest websites and prevent such pages from search results, reducing schemes that impersonate airline customer service providers by over 80 % and those that mimic standard resources like visas and government services by over 70 % in 2024.
Google announced the release of a new warnings feature for Chrome on Android that uses an on-device machine learning model to warn users of unwanted notifications that malicious websites are attempting to trick them into downloading suspicious software or providing sensitive data.
Hannah Buonomo and Sarah Krakowiak Criel of Chrome Security ,” This new feature uses on-device machine learning to detect and warn you about potentially deceptive or spammy notifications. It gives you an extra level of control over the information displayed on your device.”
You’ll see the name of the website sending the notification, a message warning that the notification’s contents may be deceptive or spammy, and the option to unsubscribe from the site or see the flagged content when a notification is flagged by Chrome.
The features come less than two months after Google released the Messages app for Android with AI-powered scam detection features. Similar methods were by the company last year to flag scam calls.
The updates come as Google appears to be releasing an for Android 16 that, in some ways, resembles Apple’s strategy by turning off JavaScript, enabling 2G connections, and turning on a number of security features by default, including Android Safe Browsing, spam protection in Messages, and the theft detection lock.
Google has also been spotted developing a tool to stop scams that entice victims to download banking apps through phone calls, according to an earlier report from Android Authority.