WhatsApp deploying Machine Learning algorithms to fight spam

WhatsApp intends to use Machine Learning algorithms to fight spam and fake news across the messaging platform, according to a whitepaper released by Facebook. This algorithm is responsible for majority removals of more than two million accounts involved in spam every single month. WhatsApp is stepping up the game in its fight to stop the spread of fake news across the platform.

The biggest hurdle in WhatsApp’s way is that it has to identify spam and the spam messages without breaking the end-to-end encryption imposed on messages. So instead of looking at the messages themselves, the algorithm explores the user habit based upon the registration meta-data which includes the mobile number from which the user signed up, mobile device details and the IP address. 20% accounts are suspended at the time of registration due to this algorithm.

To address the problem of bots, the above approach along with some other indicators such as sending 100 messages in 10 seconds within 5 minutes of registration are taken as parameters for suspension. Additionally, if the bots or some other genuine account is sending a malicious link, it will be marked as suspicious by WhatsApp to inform other users about it.

WhatsApp also bans users which have been reported but it makes sure that a particular group is not excessively targeting the same individual. Do you think WhatsApp’s tools and support regarding users being reported are sufficient to combat spam and fake news?

Source: VentureBeat