Google has revealed how machine learning is aiding in the security and enjoyment of Chrome.
To begin with, Google claims that its latest machine learning (ML) model has enabled Chrome to detect more than twice as many phishing attacks and harmful sites.
In March, the new on-device machine learning model was released. Google claims that Chrome has detected 2.5x more threats since its release.
Aside from security, Google is planning to apply machine learning to improve the Chrome user experience.
Chrome allows users to refuse notifications from pages they aren’t interested in. Google says it plans to include an AI model in the next version of Chrome that learns when users are unlikely to give prompts based on previous interactions and silences them to reduce distractions.
The design prevents users from being interrupted but allows them to receive notifications if the ML model makes a mistake (hey, it happens!).
The following step is the expansion of a tool called Journeys, which Google introduced earlier this year.
Journeys seek to assist consumers in retracing their trips online by utilizing all of the data Google collects about users. Google says Chrome will gather together all the pages you’ve viewed on a certain topic by incorporating some ML trickery. The aim is to put the days of scrolling through our whole browser history to pick up where we left off behind us.
However, it’s the final feature that’s arguably the most interesting.
Google says that it will use ML to personalize Chrome’s toolbar in real-time based on the individual user.
“Maybe you like to read news articles in the morning – phone in one hand, cereal spoon in the other – so you share lots of links from Chrome. Or maybe voice search is more your thing, as you sneak in a few questions during your transit ride to work,” wrote Tarun Bansal, Chrome software engineer, in a blog post.
“Either way, we want to make sure Chrome is meeting you where you’re at, so in the near future, we’ll be using ML to adjust the toolbar in real-time – highlighting the action that’s most useful at that moment (e.g., share a link, voice search, etc.)
The ML-powered features for Chrome really help to show how such models are improving our security while making day-to-day experiences more enjoyable.