Google Chrome’s New On-Device Machine Learning Model Can Block 2.5 Times More Potential Phishing Attacks and Possibly Malicious Sites

This Article is written as a summay by Marktechpost Staff based on the Google article 'Building a more helpful browser with machine learning'. All Credit For This Research Goes To The Researchers of This Project. Other references post 1, post 2.

Please Don't Forget To Join Our ML Subreddit

The goal of machine learning (ML) has always been to develop more valuable products to find answers to business-related questions. Google has employed machine learning to boost the use of its products, notably Chrome, since the dawn of time. Web images have been made more accessible to those with visual impairments, and real-time captions for online films have been generated for people in noisy situations or with hearing impairments. Google has made Chrome safer and more accessible due to recent technological advancements while also providing a more tailored browsing experience to its customers. Google values its customers’ privacy above all else, which is why the latest improvements are powered by on-device ML models, ensuring that users’ data remains private and never leaves their device. Every day, Chrome’s Safe Browsing protects billions of devices by displaying warnings when users attempt to visit risky sites or download dangerous files. Compared to the previous model, Google just announced a new machine learning model that can identify 2.5 times more potentially harmful sites and phishing attempts, resulting in a safer and more secure web.

The importance of online notifications and how people interact with them was also highlighted. Page notifications allow users to receive updates from the sites they visit regularly. Notification permission requests, on the other hand, might quickly become annoying. Chrome predicts when permission prompts are unlikely to be given based on how the user previously behaved with similar permission prompts and silences these unwanted prompts to enable consumers to browse the web with minimal interruption. Users can expect an ML model that produces these predictions on-device in Chrome’s future release. Google also developed Journeys to help individuals retrace their online steps by pulling together all the pages they have viewed on a specific topic and making it simple to pick up where they left off. There is also ongoing research being done to make those websites available in the language of the user’s choice. This will be accomplished by using a language identification model to determine the page’s language and whether it needs to be translated to reflect the user’s choices. This has resulted in tens of millions of successful translations per day.


Google aims to create a browser that is really and consistently useful. With the immense possibilities that come with machine learning, there are numerous untapped domains that academics are eager to study more. Continuous research is being done to enable Chrome to become a browser designed specifically for people to meet all of their demands. Future work will include applying machine learning to update the toolbar in real-time by highlighting the most practical action at the time. Link sharing, voice search, and other customizations are among them.

­čÉŁ Join the Fastest Growing AI Research Newsletter Read by Researchers from Google + NVIDIA + Meta + Stanford + MIT + Microsoft and many others...