WhatsApp is reportedly using machine learning to identify "troublemakers" before they even get accounts. Overall, about two million accounts are banned on the platform each month.
VentureBeat reports that the system was built to find users who manipulate the platform to send bulk messages or create multiple accounts to spread questionable content. Automated fake accounts and troublemakers are being barred from the platform at various stages: at point of registration, while messaging and after being reported by others.