The “hate raids” on Twitch spotlight an important online safety reality: It’s increasingly a cross-platform effort.
Hate raids, or hate spam, is harassment on steroids and, as the Washington Post points out, it’s a form of online harm that has been around for years but “became a larger Twitch concern [last] month after a streamer … posted a video of themselves getting raided.”
Because everything good online seems to have its bad side (and vice versa), raids are the downside of a good thing. Where Twitch is concerned, they’re a weaponized “platform mechanic created to amplify good behavior,” tech analyst Casey Newton explains. Popular livestreamers use it essentially to ”donate their livestream viewers to smaller channels at the end of their broadcasts, as a way of promoting up-and-coming talent.”
Bots play a big role
Bots amplify the effect, for good and bad. The Verge explains that, in addition to the “organic audience” (human viewers), the new “up and coming” streamer’s chat gets “flooded with hateful messages” from fake (non-human) “bot-generated accounts,” which makes it really hard for the streamers and their content moderators to keep things civil. Twitch says it’s working to get livestreamers tools to stop the hate raids, and so are streamers themselves (see the Verge article to read about what they’ve developed).
Trolls and raiders have workarounds that challenge platforms’ tools for stopping huge numbers of duplicate files. Whether they’re videos or text messages, if hundreds or thousands of copies are shared, software can detect and block exact copies. The workaround haters use is simple. Newton offers an example. When a racist, homophobic or misogynist message is banned, raiders throw in “lookalike characters from the
the Cyrillic or Hebrew alphabets” making the messages “just different enough to escape detection.”
Which brings us to the cross-platform issue: the fact that Twitch, for example, is only where the raid shows up, but it can be coordinated anywhere. “The damage might be done on Twitch, but it’s organized on chat platforms like Discord and signal-boosted on YouTube and far-right-friendly video sites like BitChute,” the Washington Post reports (check out the Post article to learn about streamers’ own countermeasures both technical and social, such as #ADayOffTwitch).
Clearly, cross-platform harm needs to be met with cross-platform solution development. The platforms are working hard to address viral hate in their own spaces. Twitter’s testing a new anti-harassment tool called “Safety Mode,” TechCrunch reports, and Instagram introduced “Limits” to “curb spikes in abusive messages,” Mashable reports. And the platforms work together to combat child sexual exploitation and terrorist content. Now we need to see them coordinate on countering hate speech at scale. Maybe they are and just haven’t gone public about it. But if so, streamers and users need to know.
- Added later in the month: “Twitch Sues Users Over Alleged ‘Hate Raids’ Against Streamers,” goes the Wired headline. “The lawsuit accuses two anonymous users of ‘targeting black and LGBTQIA+ streamers with racist, homophobic, sexist and other harassing content’ in violation of its terms of service,” Wired reports. The lawsuit, filed in northern California, is a sign of how hard it is to stop determined hate raiders “Stomping out botters is a bit like playing whack-a-mole,” according to Wired. The suit is aimed at revealing the identity of the two “botters,” who Twitch believes are in Austria and the Netherlands, “so they can face legal consequences.”
- On #TwitchDoBetter and other powerful hashtags: “#ADayOffTwitch and dozens of subreddit blackouts were organized separately, but together they show the potential power of collective action by users,” reports Wired.
- GameRant.com on the Twitch-YouTube dynamic
- Still, GameIndustry.biz reports that Twitch saw 27% year-over-year growth in viewership last month.”People watched 1.9 billion hours of video on Twitch in August,” they add, while Facebook Gaming reached a new record of 567 million hours watched last month.