Twitch Uses AI Tool to Catch Channel Ban Evaders

Twitch announced on Tuesday that it is developing a technology to detect individuals who try to access channels where they have been banned.  Suspicious User Detection is a new feature that will alert “likely” or “potential” channel ban evaders so that channel creators and moderators can determine whether to take action.

Users have pressed Twitch to address abuses on the platform, such as targeted harassment and so-called hate raids, in which streamers’ conversations are swamped with harassing messages. The video gaming platform said it was developing the tool in response to user input about new ways to combat ban evaders.

Twitch has been detecting specific types of user activity and account attributes to flag potential ban evaders for several months.  According to Twitch, the feature will be switched on by default for all channels, but it may be tweaked or turned off. It also stated that consumers could manually add users to the program that they want to keep an eye on.

Alison Huffman, director of product for community health, said it addresses that core problem of targeted harassment. They have heard through talking with creators and their moderators over time that this is a persistent problem for them. Twitch added a way for creators and moderators to require phone and email verification before sending chat messages, in a new anti-harassment measure. That month, it also filed a complaint in U.S. federal court against two alleged hate raiders.