The tournament platform FACEIT spoke about the success of the Minerva moderation system, developed on the basis of artificial intelligence. For 1,5 months, AI banned more than 20 thousand players.
The system was developed jointly with Jingsaw using Google Cloud. Minerva fixes violations after the end of the match. It punishes players for spamming, insults, cheating, and more. AI trained for several months using machine learning, after which it was launched on the platform.
The developers recorded the positive impact of Minerva on communication between players: after six weeks of work, the number of toxic messages decreased by 20% (from 2,2 to 1,82 million pieces). In terms of users, the number of players who allow themselves to write offensive messages has decreased by about 8%.
FACEIT has promised to develop and improve the work of Minerva in the future. The company plans to improve it so that it recognizes insults in real time.
In early September, Kaspersky Lab
Source: 3dnews.ru