Twitch has introduced a new tool to help fight harassment in livestream chats.
Suspicious User Detection is a new tool that allows creators and mods to catch users attempting to evade channel-level bans.
“When you ban someone from your channel, they should be banned from your community for good,” said Twitch in a blog post. “Unfortunately, bad actors often choose to create new accounts, jump back into Chat, and continue their abusive behavior.”
The tool is designed to prevent users from simply creating new accounts when banned for their abusive behaviour. Powered by machine learning, Suspicious User Detection will identify these users based on a number of account signals.
Suspicious User Detection will flag suspicious accounts as either “likely” or “possible” ban-evaders, so that creators and mods can take action when needed.
Messages from “likely” ban evaders won’t appear in chat at all – though their messages will be visible to creators and mods, who can choose to leave the restrictions in place, remove them or ban them from the channel entirely.
“Possible” ban evaders meanwhile will have their messages appear in chat as normal, but the account will be flagged to the creator and mods, so that their chat can be restricted if needed.
The new feature will be enabled by default, and can be disabled or adjusted via the channel’s moderation settings page. Potentially abusive users can be monitored more closely via the new Suspicious Users widget.
Twitch notes, however, that no machine learning will ever be 100% accurate,” and so there is a possibility of false positives and false negatives – which is why the new tool doesn’t automatically ban users.
“This tool highlights our overall approach to safety technology: build powerful tools that work together to give you finer control over your community,” said Twitch. “Our work will never be finished and we’re continuing to develop more tools to prevent hate, harassment, and ban evasion on Twitch”