As hate raids continue to plague Twitch, with marginalized streamers targeted, one has exposed just how difficult it can be to police chat, despite the platform giving creators tools to do so.
Hate raids are a very concerning topic in the Twitch community, with armies of bots designed with one purpose: spam racist and other offensive terms in chats.
This has resulted in protests, hashtags and cries for the Amazon-owned site to take action. While Twitch has acknowledged the problem and has promised change, streamers don’t feel it’s enough, especially when the tools at their disposal have major drawbacks.
Content creators are advised by Twitch to use tools provided to protect themselves from harassment, such as adding offensive terms to a block list. However, as shown by streamer Art for the Apocalypse, comprehensively banning all variations of offensive terms is nearly impossible.
In a video, Art for the Apocalypse showcased how just trying to block the word “jogger” can take days, all because of all the variations of the word using special characters.
“Harassers on Twitch often circumvent blocked words listed by swapping out characters in their chat messages with characters from other languages that are still easily recognizable,” he explained.
Basically, in addition to banning the word “jogger” he needs to ban many variations. By creating a script, he created all the variations possible – all 21.9 million of them.
Because Twitch doesn’t have an option to upload blocked terms en masse, he needs to set up his channel bot to block the terms, but there’s a problem with that too: Twitch limits the number of messages a bot can send to 100 every 30 seconds.
This means that to ban one word and its multiple spellings using the Latin alphabet, his channel bot would need to run 24/7 for 76 days.
We're launching channel-level ban evasion detection and account verification improvements later this year. We’re working hard to launch these tools as soon as possible, and we hope they will have a big impact. Check out more on our existing tools here: https://t.co/Dku6eBhY72
— Twitch (@Twitch) August 11, 2021
“Expecting streamers to manage this problem themselves is not only unfair, it’s not technologically possible,” he slammed. “Do better, Twitch.”
Twitch has outlined some of its plans to combat hate raids, but a lot of its improvements will only be going live later in the year. Hopefully, something can be done sooner to give streamers the power to ban certain words from their streams a lot faster.