Google Boots Up AI Neural Network To Rid Your Inbox Of ‘Sneaky Spam’

Gmail reduced spam from a major headache to an occasional annoyance, but Google is going after the remaining 0.1 percent of garbage that makes it to your inbox. The company rolled out new tools this week that are meant to shut down the most pernicious spam strategies and help important emails avoid getting inadvertently trapped in the spam filter.

Google product manager Sri Harsha Somanchi describes the spam that makes it past the company’s filter as “Sneaky spam – the kind that could actually pass for wanted mail.” To nail it, Google is now employing an artificial neural network, which is machine-learning technology Google has had under development for years. Gmail should also be better at determining whether a message has actually come from the supposed sender.

Google gmail

Interestingly, Google’s machine learning isn’t focused solely on the bad guys. Gmail also aims its learning capabilities at you, using your behavior to determine whether you like to receive certain types of legitimate email. If you’re not a big fan of weekly newsletters, for example, Gmail should figure that out and cull them as they arrive.

Companies that send legitimate emails in bulk are getting new Gmail Postmaster Tools, which should provide some protection from Gmail’s spam filter. Google estimates the amount of legitimate emails landing in your spam box at about 0.05 percent, but it’s aiming for even better accuracy. That’s welcome news for anyone who’s missed an important email thanks to the spam filter. The Postmaster Tools will help these companies analyze and improve their emailing practices. “For you, this means no more dumpster diving for that confirmation code,” Somanchi said in a blog post.

For most of us, spam isn’t the problem that it once was (thanks in large part to email services), but it’s good to see that Google isn’t kicking back.