Google uses AI to rid your Gmail inbox of more spam

Spam is always annoying, but it can occasionally be disastrous. Google has now deployed its artificial neural network to stop more of it from arriving in your Gmail inbox, something it hinted at earlier. It's designed to "detect and block the especially sneaky spam -- the kind that could actually pass for wanted mail," according to the company. The system also uses machine learning to track your usage patterns and figure out if you want certain kinds of mail, like newsletters or promos. Most critically, Google said that Gmail is now better at catching impersonation -- when emails appear to be from a known contact, but were sent by someone who is definitely not your friend.

Google is also working the other side of the spam equation with its new Postmaster Tools, aimed at high-volume senders. Qualified companies who meet Google's reputation requirement will get access to services that show how Gmail handles their emails. For instance, a company that sends out tons of legitimate mail -- like order or flight confirmations -- will see how often users mark it as spam. They can then tweak it so that customers treat it more seriously. From an end-user perspective, Google said the new tools will make it less likely you'll need to go "dumpster diving" in spam to find important messages.