New Twitch Tool Is Coming After Toxic People

Twitch is cracking down on toxic users with a new tool aimed at reducing the potential abuse flowing streamers' chat. A far cry from the platform's recent Boost feature causing an uproar and adding a pay-to-win element to streaming, the new Suspicious User Detection tool seems like it's genuinely there to help. The feature is powered by machine learning, which can teach itself new techniques over time. By recognizing any suspicious behaviors in a streamer's chat, it hopes to catch those who would try to circumvent bans (per Eurogamer). Twitch has a new tool at its disposal, and the platform isn't afraid to use it to weed out toxic people.

Suspicious User Detection comes after a reported increase in harassment and hate raids on streamers belonging to minority or marginalized groups. In a blog post announcing the feature, Twitch said Suspicious User Detection is there "to help Creators and Mods catch and take action against users who are trying to evade channel-level bans." Twitch's new feature sounds like a great addition to any streamer's utility belt, giving them more control over their chats. Suspicious User Detection will be turned on by default for all channels, but here's how the feature actually works and how it can be altered or disabled.

In its blog, Twitch acknowledged that problem users are often able to bypass bans or restrictions by creating new accounts and then simply resume the harassment. To combat this, Suspicious User Detection constantly monitors "signals" emitted by an account. Presumably, these "signals" have to do with things like IP addresses, account creation, or activity trends and other identifiers that could help paint a target on the back of a bad actor. Though Twitch hasn't thoroughly explained the "signals," they confirmed the tool "will flag suspicious accounts as either 'likely' or 'possible' channel-ban evaders," allowing Creators or Mods to take action if needed.

There are two potential states for accounts flagged by Suspicious User Detection: "likely" and "possible." If a user is marked a "likely" channel-ban evader, their messages won't be seen throughout the chat, only to Mods and Creators. For those that are considered "possible" offenders, their messages will appear in chat as usual, but the account will appear flagged to Mods and Creators, so the user can easily be monitored or banned. Though the feature is enabled by default, it can be deactivated or further customized from the moderation settings page.

Though many have criticized Twitch's response to hate raids and negative behavior, Suspicious User Detection seems like a meaningful step by the company to address these issues. Hopefully, Twitch's new feature should help make the platform a bit safer.