the Company Faceit with Google Cloud and Jigsaw wrote bot with artificial intelligence that blocks the toxic players in Counter-Strike: Global Offensive.

According to the company, they are a few months “coached” bot administrator Minerva using machine learning to “minimize false positives”. If the bot finds a message that in the context of the conversation is assessed as “toxic”, it sends its author a warning, and all such reports will mark as “spam”. The punishment becomes more severe if the player commits a second violation for a short period of time. Minerva is able to make decisions in just a few seconds after the end of the match and when it detects the intruder immediately sends him a warning message or announcement on the bath.

Faceit launched the bot in automatic mode at the end of August. A month and a half of work, he analyzed 200 million emails, of which 7 million he was assessed as toxic. During this time, Minerva “made” 90 thousand warnings and 20 thousand lei for abuse and spam. According to the company, the amount of toxic posts for the first month of operation decreased by 20.13% from 2,280,769 in August to 1,821,723 in September.