Google has announced that it has deployed a new AI-powered spam protection measure for its flagship Gmail webmail service.
This new spam protection is being powered by the Google-developed open source TensorFlow machine learning framework.
In the context of the over 1.5 billion monthly users of Gmail, an additional 100 million spam messages a day being blocked may not seem like much on paper, but Google's previously existing protections against spam were purported to block 99.9 percent of it. This additional 100 million spam emails knocks out a huge chunk of the remaining 0.1 percent, even if it only works out to a single additional blocked spam email for every ten users.
Google is currently looking into using this TensorFlow-powered spam protection in other ways, such as protection against malware and phishing attempts, according to the blog post announcing TensorFlow's Gmail spam protection. Nothing concrete has been announced, however.