Chances are, you have more emails in your spam folder than your inbox on Gmail. Nowadays, spam messages are becoming more and more frequent, starting from annoying newsletters and fake job invitations to a Saudi prince being stuck at an airport. However, Google wants to put that to an end with its new AI, which will block 100 million extra Gmail spam messages per day.
Spam messages may go to a flagged folder, but that doesn’t make them any less annoying. Google’s AI framework TensorFlow will be trained to additionally filter Gmail inboxes, the company announced. The framework has been deployed for a month now, with Google claiming it can block 100 million extra spam messages on a daily basis.
“At the scale we’re operating at, an additional 100 million is not easy to come by,” Neil Kumaran, product manager of Counter Abuse Technology at Google, told The Verge. “Getting the last bit of incremental spam is increasingly hard, [but] TensorFlow has been great for closing that gap.”
Since Google has already been using AI for a long time to block spam messages, this is not huge news. However, current filtering techniques can block only obvious spam, while the machine learning system will learn from new patterns which hint at spam. More importantly, it will determine whether someone’s Gmail is safe or not.
According to Kumaran, TensorFlow manages data on a much easier level. Also, it’s been open source since 2015, meaning that new innovations and research from the community can be quickly put to work.
Launched since 2015 by Google, TensorFlow is one of the most important AI resources for the search engine giant. Being free, developers can develop a huge variety of machine learning programs which can complete various tasks. People that use it are praising its flexibility and ease of use, while also being easy to integrate with other AI services that the company offers.
Now that Google has integrated TensorFlow into its messaging system, allowing it to block 100 million extra gmail spam messages a day, Gmail will have a more personalized filtering of spam messages, while keeping messages that may appeal to that particular user. Google has been trying to implement this technique for many years, according to Kumaran. However, the company needed users to signal what they found to be more spammy among the sea of emails. Now with TensorFlow, there will be much better results.
“There’s no one definition of spam out there,” he adds. But AI would help work out the best protection for each individual.
In addition to more security to Gmail, Google has also recently released a new Chrome plugin which recognizes when a password is exposed to risk.