Visitors walk past televisions with live streams from Twitch Interactive’s video service at the E3 Electronic Entertainment Expo in Los Angeles.
Patrick T. Fallon | Bloomberg | Getty Images
Twitch, Amazon’s video streaming platform primarily used by gamers to live stream their games, announced a new policy on Wednesday that allows the company to take action against users who engage in certain harmful behavior completely offline.
The policy represents a unique approach from peers on social media at a time when the industry is under increasing pressure to implement strong and consistent content moderation policies. As legislators on both sides of the aisle have threatened to strip online platforms of their liability protections under section 230 of the Communications Decency Act, many platforms have taken steps to put stronger guardrails in front of what users can post.
Under the new policy, Twitch can suspend users indefinitely after a third-party researcher determines that there is strong evidence that the person engaged in certain offline behavior. These actions include committing deadly violence, terrorist activities, grooming children for sexual exploitation, committing sexual assault, or even “acting as an accomplice in unconscious sexual activities”. It will also continue to consider offline harassment in cases where a user alleges abuse online.
Twitch said it will work with “an experienced investigation law firm” to determine the validity of claims, which will sometimes rely on access to law enforcement evidence. The company said it would not take action on a user’s account until it completed its investigation and confirmed evidence of wrongdoing.
The malicious offline behavior does not have to involve another Twitch user to be considered a violation, a spokesperson confirmed. That’s based on the idea that people who engage in this type of behavior are more likely to create security risks for the Twitch community, the spokesperson added.
Other social media platforms also consider real harm propagated by users on their platforms, but Twitch’s new policy is unique for its explicit approach to completely offline behavior and for some types of offline behavior it prohibits. For example, Facebook’s community standards prevent mass murderers and members of terrorist, hate, criminal or human trafficking organizations from being present in any way on its platforms. Twitch’s policies include other offline behavior that may not be part of an organized crime group, such as committing sexual assault.
Social media platforms base the majority of their enforcement actions on damage resulting from content actually posted on their services. While they can take real world events into account when evaluating the damage, they will usually refer to posts on their own platforms as the tipping point for action.
Even when Facebook and Twitter each decided to ban former President Donald Trump from their services following the January 6 uprising at the Capitol, their reasoning was fundamentally rooted in how they said he was using or potentially using their platforms to incite further violence. .
Twitch’s new approach comes after a broader discussion of how certain real-world events should be handled by the tech platforms. Last month, a Business Insider investigation revealed a woman’s accusation against Dominykas Zeglaitis, a member of the so-called Vlog squad led by popular vlogger David Dobrik. The unnamed woman said that Zeglaitis sexually assaulted her one night she and her friends appeared in one of the group’s videos when she said she was too drunk to consent. Zeglaitis declined to comment on the allegations made to Insider.
Google-owned YouTube said after the report it would temporarily prevent Dobrik from monetizing his account through ads. Under YouTube’s Creator Policy, off-platform behavior, including violence or cruelty, can lead to penalties, such as missing out on promotional opportunities or having their videos appear in recommendations to users.
While Twitch will initially address a handful of said serious violations, the platform said it strives for the guidelines to be iterative. Because offline damage is difficult to verify, the company has prioritized categories that the company believes are most harmful to the community first.
Users who wish to confidentially report offline damages that fall under the categories prohibited by Twitch can send an email to [email protected].
Subscribe to CNBC on YouTube.
WATCH: The Big Messy Case of Content Moderation on Facebook, Twitter, YouTube