Twitch is a place of great fun, but also a place of great trolling, spam and harassment. Sexism, racism and general spite is ever-present, and as messages are often scrolling by at an unobservable speed in the more popular channels it can be an uphill battle when it comes to putting an end to the hate-speech.
Even the most skilled, just and observant human moderators aren't able to keep up with ten thousand typing viewers, even when only one percent of them are trolling or intentionally sabotaging the channel. There's just too much information to sort through—for a human, at least.
Luckily, we're living in 2016 and we still control our robot creations. Streaming giant Twitch has introduced it's answer to the vast hordes of chatroom trolls, called AutoMod.
Filtering Out The Trash Talk
So how does AutoMod actually work? Well, as it is now, most harmful messages on Twitch go live and are readable before a human moderator has a chance to remove them and block the spiteful user. In these cases, the damage has already been done. AutoMod, on the other hand, circumvents this and actually filters messages preemptively, before they're even published.
AutoMod judges comments using four different moderation categories: identity, sexual language, aggressive speech and profanity. The mod checks the chat messages against its filters and either approves or dismisses them. If a message is dismissed, AutoMod will send it to a moderation queue where it'll be reviewed by a human moderator.
Will The Salty Trolls Find Another Way?
Twitch tested AutoMod during the Democratic and Republican National Conventions earlier this year, and will now offer the system to all users. Some may find the mod as a bit of an over-the-top reaction, but the issue of bigot chats has undoubtedly caused problems in the past. And if a streamer still wishes to leave their chat unfettered, they can do so, as AutoMod is completely optional.
The more persistent trolls of Twitch will surely try to find ways around the new filters, but according to Twitch's press release AutoMod isn't just a machine that bans words:
"AutoMod employs machine learning and natural language processing to identify and block inappropriate content from appearing in chat. Beyond identifying inappropriate words and phrases, AutoMod can detect potentially inappropriate strings of emotes and other characters or symbols that others could use to evade filtering."
We'll have to wait and see how the great battle between AutoMod and the Twitch trolls will play out.
What are your experiences with the Twitch chat? Do you think AutoMod is a necessary measure?