Gmail Now Uses Artificial Neural Networks to Sniff Out Spam

By Jamie Condliffe on at

Gmail’s spam filters have always been pretty good, but now they’re getting a shot in the arm. Google’s rolling out its artificial neural network technology, currently used in the likes of its Search and Now apps, to help reduce the weight of unwanted email even further.

The filters that Google has used in the past have been based around rules written by engineers, but now they’ll be able to learn for themselves. Using information garnered from your clicking of “Report spam” and “Not spam” buttons, the networks will help learn themselves what constitutes spam and what doesn’t. In a blog post, Google explains that the new feature will “detect and block the especially sneaky spam”.

The AI will also learn about your individual preferences, too. “While your neighbour may love weekly email newsletters, you may loathe them,” writes Google. “The spam filter can now reflect these individual preferences.” Finally, the filters will also be able to do a better job of spotting impersonation (the kind used for phishing attacks) to ensure you’re only shown genuine email. All of which means your Gmail inbox should be a little clearer from now on. [Google via Engadget]

Image by Cairo under Creative Commons licence.