Google and WhatsApp are using AI to kick users, flag spam on a massive scale
Google and WhatsApp have each shared how they are using their own artificial intelligence engines to cut down on bad actors in their Gmail and messaging services.
On the Google Cloud blog, the company announced it has started to implement TensorFlow, its dedicated framework, in addition to existing machine learning and rules-based guides to detect spam messages on Gmail. AI tools are able to scrutinize what users are deleting and act on that information for incoming messages.
“Just because some of an email’s characteristics match up to those commonly considered ‘spammy,’ doesn’t necessarily mean it’s spam,” wrote Neil Kumaran, product manager of counter-abuse technology at Google.
While the company estimates that its basic system was able to label spam with 99.9 percent accuracy, it found that it could grab about 100 million additional messages a day.
Meanwhile in India, WhatsApp held a media briefing to publicize its efforts to cut down on fake news as another general election looms. A month of voting is scheduled to begin in April and given that waves of fake news stories shared on the Facebook-owned platform have generated quick outrage and outbursts of violence, those efforts are more important than ever.
VentureBeat reports the company revealed that it has unleashed a machine learning system to track down bad actors and ban them from the network. 2 million accounts are being dumped each month and 20 percent of those dumpings happened even before habitual users created new accounts.
It has also limited the number of text forwards and highlighted fact-checking media to enhance safeguards against fake news.