Twitch recently launched a new tool that i powered by AI to help its users identify those who were trying to bypass channel bans. The platform which is mostly used for streaming games stated that the new tool was part of an ongoing mission to keep the sections of chats of user channels completely free from harassment or hate.
While most users can already ban people permanently from their chat, the streaming platform mentioned that people would often try to cheat the ban by creating new accounts so they can get back to the chat. This new app uses the Suspicious User Detection tool which analyses a number of account signals using machine learning, which would help spot users who have been trying to dodge their bans by creating new accounts.
Twitch’s New Tool Would Keep Chats Clean
This Twitch tool would then flag the suspicious accounts, either as possible or likely channel ban evaders, after which the moderators and the owner of the channel would be informed. If an account is flagged as likely, the messages from the account wouldn’t be sent to the chat, and will only be seen by the moderators as well as the creators- who can obviously choose to leave the restriction in place- while monitoring the user, or banning them from the channel.
If the Twitch tool sees the account as a possible ban-evader, the message would simply go into chat normally, but then the account would be flagged for moderators and creators- allowing them to monitor the user while restricting them from chatting- if that is seen as necessary. The tool is activated automatically for channels, and streamers would have the opportunity to disable it if they so wanted. Users would also be able to manually add accounts that they think need to be looked at.
Twitch had noted that machine learning was never completely accurate, which is precisely why the Suspicious User Detection tool outrightly bans accounts that it flags.
