Twitch has just released a tool that uses machine learning to detect users who try to re-enter chats from which they have been banned for abusive behavior.
The gaming live streaming platform said that already banned users once tried to bypass Twitch’s controls by creating new accounts to continue harassing other users.
But the new system released a few now manages to warn streamers and chat moderators if a particular user is a “probable” or “possible” ban bypasser .
This new technology employed is part of Twitch’s long-term strategy to reduce the climate of hatred and harassment that all too often arise during live broadcasts.
The company was criticized for “hate speech” when it was discovered that some unscrupulous streamers were sending their followers or even automated bots on other channels to harass other users.
Often the victims belong to minorities or marginalized groups.
The creators had asked the Amazon-owned company to take a few more steps to counter this type of hate speech.
In September 2021, Twitch announced “phone verified chat , ” allowing streamers to require some or all users to verify a phone number before chatting .
And the same month, he filed a lawsuit against unidentified users, allegedly involved in “chat-based attacks against marginalized streamers.”