Twitch announced Wednesday that it is rolling out phone-verified chats, a safety feature to help get harassment, malicious users and so-called hate raids under control.
Twitch has come under close scrutiny of how it handles the safety of its Black and LGBTQ users. "Hate raids" — when users or bots infiltrate chats with harassment — have made streaming on the platform untenable for many users who belong to marginalized groups.
Phone-verified chats will allow targeted creators to have more control over who can participate in their communities, Twitch said in a blogpost Wednesday.
The move will require all or some users to verify phone their numbers before they can participate in streamers' chats, which could help reduce targeted attacks by bots. Once users verify the phone numbers to their accounts, they will not need to verify them again for other channels.
Twitch said it began working on the tool five months ago, before the recent spotlight on targeted hate raids, and it said it is optimistic that the new tool could reduce harassment and bot attacks.
Twitch's vice president of trust and safety, Angela Hession, said that the new tool was not designed exclusively in response to hate raids; instead, she said, it is "intended to help stop chat-based harassment of all kinds."
"Hate and harassment of any kind is unacceptable on Twitch, and we know we have further to go in building the safe, inclusive community we intend," Hession said by email.
Hession said the bot attacks were frustrating, adding that Twitch is working to make the platform a safer and friendlier place for all its users.
To reduce harassment and other concerns, Twitch has doubled its safety operations team in the last year and quadrupled its number of moderators, Hession said. She said the increase in moderators raised the team's response time to users by 96 percent.
"Creators have different needs, and we've learned that we can best create a safe, diverse environment at scale when we empower creators to set their own guidelines on top of our own," she said.
Twitch said users will be able to customize who must verify or have verified their phone numbers, email addresses or both. Verification can apply to age groups or time frames, such as how long ago an account was created, Hession said.
The move is also intended to prevent users from being able to evade bans by simply creating new accounts, as all accounts associated with a phone number will be banned if one account is banned. Twitch allows one phone number to associate up to five accounts with it.
Although live user attacks are a problem, bot attacks are a ferocious problem without a simple solution. Twitch said it is working to fight the issue from several angles.
"We've banned millions of malicious follow attempts and removed thousands of malicious bot accounts, but these actors have been relentless and their attacks don't have a quick fix," Hession said.
Twitch filed a complaint in U.S. District Court for Northern California this month against two of its users for repeatedly flouting its community guidelines against harassment, including committing bot attacks.
Twitch accuses Cruzzcontrol, a streamer believed to be based in Baarto, Netherlands, and Creatineoverdose, a streamer believed to be based in Vienna, of engaging in hate raids, creating new accounts to continue harassing minority Twitch users after the platform banned them and using bots to overwhelm creators with sexist, racist and homophobic language and content, according to the court documents.
The two streamers are known only by their usernames.
Thousands of bots have been linked to Cruzzcontrol, "including those targeting black and LGBTQIA+ streamers with racist, homophobic, sexist and other harassing content," according to the complaint. Creatineoverdose also used bot software to harass marginalized users.
Ahead of the suit, marginalized creators held a boycott, which they called "A Day Off Twitch," to raise attention for how rampant hate raids had become. A spokesperson said in a statement at the time that Twitch was "working hard on improved channel-level ban evasion detection and additional account improvements to help make Twitch a safer place for creators."
Hession was reluctant to share details about what Twitch plans next for safety, saying such information could tip off bad actors and give them head starts on circumventing the tools.
"I know it's frustrating when we can't share details about our safety pipeline, especially when community members are hurting—but we are always working on improvements and actively incorporating community feedback into the work we do," she said. "And there is more to come."
Some Twitch streamers, who have been the targets of hate raids themselves, said they were glad the platform is prioritizing their safety.
"It's not perfect, and there is still a long line of other changes that are necessary to make Twitch safe again," said Lucia Everblack, a streamer who helped to organize the boycott. "But it has me feeling hopeful."