Google’s Gmail generally does a respectable job in fishing out spam but that certainly doesn’t stop the tech giant from turning to AI to make spam blocking even more effective.
Google in a recent blog post mentions that the company has been using new AI-driven filters over the past month which has enabled Gmail to block an additional 100 million spam message every day. This is in addition to the 99.9 percent of spam messages Google already claims Gmail blocks.
Google has recruited its open in-house machine learning framework, TensorFlow to further train Gmails algorithms and detect spam from the likes of newly created domains, image-based messages, and even messages with hidden embedded content.
While 100 million spam messages every day may sound an immense number, as pointed out by The Verge, this isn’t really as significant a number as Google makes it seem. As per Google’s own estimation, there are currently 1.5 billion users who use the company’s mail service. Spreading out 100 million messages over 1.5 billion users really only gets you roughly “one extra blocked spam email per 10 users.”
But that statistic doesn’t certainly demean the impact of TensorFlow on blocking spam for Google. Getting its machine learning algorithms to work in tandem with Gmail’s existing rule-based filters and then block an additional 100 million emails everyday is still an important feat.
Moreover, this only suggests how much better Google’s TensorFlow will get over time, ensuring spam goes right where it belongs, even without you seeing it.