Introducing Our New Hateful Conduct & Harassment Policy
Every day, millions of people come together on Twitch to share their passion and skill, support their favorite streamers, and connect with friends. Communities on Twitch are diverse and are growing rapidly, representing nearly every interest, background, and language around the globe. As we evolve, it’s necessary to revisit and update policies to make sure all streamers and all communities can create together safely on Twitch.
Today, we’re introducing an updated Hateful Conduct and Harassment Policy. This new policy will take effect on January 22, 2021. Content created on or after January 22 will be evaluated under the new guidelines.
These changes are intended to better protect the community, not to be punitive, so we are sharing them before the updated policy goes into effect to ensure you have adequate time to understand what is changing. During this education period, our Safety team will continue to moderate content using the existing guidelines. We understand that any change in policy requires you to invest time to read the new language and consider how it might impact your behavior, and we are grateful for your support as we work together to make Twitch safer for everyone.
We know that many people on Twitch–particularly women, members of the LGBTQIA+ community, Black, Indigenous, and people of color–unfortunately continue to experience a disproportionate amount of harassment and abuse online, including on our service. Not only is this blatantly unacceptable, it also undermines the community we’re building on Twitch and threatens the long term viability of streaming as a career for everyone who wishes to pursue it.
We’re incredibly proud of our Creators who show up every day to create great content and build welcoming spaces for their communities, and facing harassment because of race, gender, or any other protected characteristic is unacceptable and has no place on Twitch.
We developed the new policy to take a clearer and more consistent stance against hate and harassment, and to give you greater insight on what is and isn’t acceptable on Twitch. Hateful conduct and harassment have always been prohibited, but we’ve added more detail explaining the behaviors that fall into these categories. We’ve also added specific examples to help you understand how the policy is applied in practice.
This policy is the culmination of a months-long process incorporating extensive research within our community and consultation with industry experts across fields, as well as our Safety Advisory Council. We also drew heavily on a review of past cases to identify edge cases, and opportunities for clarification. Understanding how we’ve historically enforced these policies helped us address common points of confusion and increase enforcement consistency.
What’s New in the Policy
To ensure we’ve provided adequate detail, we’ve broken the policy out into three sections: Harassment, Hateful Conduct, and Sexual Harassment, and we’ll walk through key takeaways for each.
Words and actions have meaning and impact, even if your intent is not meant to be hurtful or cause harm. And even if the target of your behavior or comments isn’t bothered by them, others in the community may nevertheless feel impacted by it. Under the new policy, our Safety team will look at the content of statements or actions in order to determine whether a behavior is abusive and violates our guidelines, rather than relying solely on perceived intent.
As a consequence of this, our assessment of some areas, including some types of harassment, may rely on indications that a behavior is unwanted, including channel time-outs, bans or reports from the targeted user or a mod team. This will provide us with the necessary context to ensure we aren’t punishing people for good natured or competitive banter that is accepted in some communities. That said, we’ll continue to weigh both the perspective of the reporter and the content under review to reach a determination. We expect these changes to improve overall enforcement consistency as they reduce subjectivity in the review process.
Harassment encompasses targeted attacks through attention seeking behaviors that are designed to disrupt, harm and or hijack a community or stream. People experience harassment in many ways, including personal attacks, promotion of physical harm, and malicious brigading. We do not allow harassment on Twitch and will take action to remove it from our service. We do want to enable users to express themselves naturally with their friends and communities without fear that these interactions could be misidentified as harassment. If you feel that a behavior crosses the line into abuse, we encourage you to use our report feature or take channel-level action to indicate that a behavior is unwanted so that we have the context we need to take appropriate action. We take your reports seriously, and each report is sent to a member of our Safety team for review.
Beyond that, several of the changes in this updated policy make certain aspects of our longstanding policy clearer and more explicit. For example, the following behaviors are expressly prohibited:
- Claiming that the victim of a well-documented violent tragedy is a crisis actor, or is lying
- Encouraging others to DDoS, hack, doxx, or swat another person
- Inciting malicious raids of another person’s social media profiles off Twitch
Harassment becomes hateful conduct when the attack is based on identity characteristics. We consider the following identity-based attributes to be protected characteristics: race, ethnicity, color, caste, national origin, immigration status, religion, sex, gender, gender identity, sexual orientation, disability, serious medical condition, and veteran status. We’ve added caste, color, and immigration status to this list to ensure we are evolving with our global community and providing sufficient protection for under-represented groups.
We’ve included more detail so you can better understand which behaviors would cross the line. For example, hate speech and symbols have always been prohibited, and we’ve added explicit language banning hate groups, membership in hate groups, and sharing of hate group propaganda. We’ve also specifically prohibited black/brown/yellow/redface unless they are being used in an explicitly educational context. Again this is not new, but the new guidelines will make the standard clearer for everyone. For more examples of prohibited behaviors, read our Learn More article on Hateful Conduct.
The following changes should also be noted:
- Emotes are an important part of how we communicate with one another on Twitch, but they can be used maliciously. So emote combinations, even without additional text used in chat, will be held to this policy
- Given its historic and symbolic association with slavery and white supremacist groups in the US, displaying the Confederate flag is prohibited
While sexual harassment has always been prohibited on Twitch, we’ve heard from our community that our guidelines didn’t adequately account for harassing behaviors that are sexual in nature. With this update we’ve separated sexual harassment into its own category and adopted a much lower tolerance for objectifying or harassing behavior.
Notable changes include:
- Repeatedly commenting on someone’s perceived attractiveness, even in what you believe to be a positive or complimentary manner, is prohibited if there is indication that it’s unwelcome (i.e. you’ve been asked to stop, timed-out, or channel-banned)
- Making lewd or explicit comments about anyone’s sexuality or physical appearance is prohibited. Note that we do not make an exception for public figures
- Sending unwanted/unsolicited links to nude images or videos is prohibited
This is another step toward recognizing our safety visions for Twitch. We do recognize we still have a lot of work to do, and we are committed to investing the necessary time and resources to get this right. We also want to acknowledge our appreciation to you, the community, for joining us in this shared commitment to make Twitch better, and for providing us with feedback along the way.
How to Learn More
We want to make sure you have the information you need and the opportunity to ask questions and get clarity before we start enforcing these new rules on January 22. So, we’ve set up three live sessions where we’ll walk you through the changes and what they mean for you, and answer your questions. If you can’t catch these live, the VODs will be available for you to check out too.
12/11 Creator Camp: Deep Dive on the Policy
10am PT on /CreatorCamp
While it’s called “Creator Camp,” the entire Community is invited! Join us for a live discussion of key policy changes. We’ll also answer questions from chat.
12/16 Town Hall: Overview of the Policy and Enforcement
10am PT on/twitch
We will discuss safety on Twitch, our goals for this new policy, and answer questions from chat.
1/20 Creator Camp: Reviewing FAQ since Launch
12pm PT on/CreatorCamp
We’ll return to the Creator Camp stage to do a rundown of frequently asked questions since the policy was announced and provide an opportunity to ask any last questions before enforcement begins.
Review the full policy language here.
Some areas of the new policy are stricter. Are you imposing increasingly strict penalties for violations under the new policy?
While we are firmly addressing some behaviors that were not explicitly included in the previous policy, this iteration of our guidelines is also much more detailed, with enforcements tailored to the severity of the action or language. As a result, behaviors that are relatively low in severity, or language commonly viewed as being colloquial, will receive warnings or lighter suspensions, while more malicious or overtly harmful behaviors and language will receive stricter penalties - the most severe violations will continue to receive an indefinite suspension on the first offense. In total, we expect that the penalties under this iteration of the policy will scale more appropriately to the severity of the violative behavior.
Will you issue warnings before the new policy enforcement begins?
Our Safety team will continue to review reports under the existing policy and issue enforcements accordingly during the period leading up to January 22nd. We will not be issuing warnings under the new policy during this period, which is why it is important that you review all the materials provided to ensure you understand what is changing. If you have questions about the new policy we suggest you tune into one of our live sessions, which will all incorporate time for Q&A. We will also be updating the blog post with additional questions from the community at various times before January 22nd.
Why are you delaying enforcement? Won’t that just allow bad behavior to persist in the meantime?
Our existing Community Guidelines already address many of the common types of behavior that are unacceptable Twitch. We issue policy updates to improve community safety, not to be punitive. Given that there’s a lot of new detail in this policy update, we felt it was essential to give the community time to read, process, and ask questions about them. This additional time will allow us to address common points of confusion and prevent Creators and their communities from accidentally running afoul. In the meantime, our safety specialists will continue to enforce our existing policy and ensure that hateful and harassing content and behavior is removed from Twitch.
How will this new policy impact older content?
This updated policy will apply only to content that is created on or after January 22, 2021. Mistakes do happen, and if you receive a strike for what you believe to be older content that was acceptable under the policy in place at the time it was created, we encourage you to file an appeal and a member of our Safety Ops team will review the content at issue.
How do you decide the length of a suspension?
While we do not make our specific enforcements public, they are based on several factors, including the severity of the violation, whether a violation is repeated, and the user’s enforcement history. Enforcements may include warnings and temporary suspensions ranging from 24hrs to 30 days. In the most severe cases, we will issue indefinite suspensions on the first offense, or due to repeat offenses.
What do you mean when you say that you will allow the use of some words or terms that might otherwise violate your policy against hateful slurs? How do you determine which words fall under this category?
We acknowledge that some slurs have a history of reclamation or colloquial use to indicate belonging to certain groups or communities. While these terms are traditionally tied to identity, we do not believe that Twitch is, or should be, in a position to determine an individual’s identity. After consultation with a number of diversity & inclusivity academics, we have decided to allow the use of some terms that may otherwise be viewed as slurs so as not to alienate or limit the expression of communities on Twitch. However, we still prohibit slurs when used in an explicitly hateful or targeted manner. This is not a perfect solution but we think it is the best option to ensure that Twitch is a truly inclusive space.
How do you determine if something is a hate symbol?
We define hate symbols as the emblems, branding, and gestures most frequently used to convey hateful ideologies or denigrate against others based on protected characteristics, including symbols of the Nazi Party, white supremacist groups, and other hate groups. We consider the context in which a symbol appears when evaluating whether it is a violation of our Community Guidelines. In particular, we allow for the presentation and discussion of symbols for the purposes of education, or in the context of historically set games, but this context must be made explicitly clear. Additionally, symbols appearing in the context of developer-generated gaming content will not result in an enforcement as long as the broadcaster does not leverage them to carry out any hateful behavior.
How do you determine if something is a slur?
A hateful slur is language that generally degrades or denigrates based on a protected characteristic. When we evaluate the use of a hateful slur, we take context into account, and we do make an exception for the use of some words or terms in an empowering way, or as terms of belonging when such intent is clear. We enforce against the use of hateful slurs globally, and in multiple languages.
How will you take action against harassment in cases where the user does not file a report?
If we receive a report that a user appears to be the target of harassment, and we have any sort of indication (moderator actions, feedback from the creator on stream), then we will take action. In the event of a more subjective situation, for example competitive banter or trash talk, will rely on signals from the person being targeted, or their mods, so that we can take appropriate action and avoid punishing users in situations where that behavior is welcome.
What do the Community Guidelines mean by “first-party reporting”?
A first party report is one submitted by someone who is the target of harmful or abusive behavior. If it is not clear that behavior is intended to be harmful, a report from the person on the receiving end of the behavior, or from one of their moderators, can help us establish that the behavior is unwanted.
Why is this policy stricter than some societal norms?
Twitch is home to communities all around the world, with different expectations and standards. Our goal is to create policies that adequately meet the needs of all those communities. Societal expectations range broadly around the world so we have developed a set of standards for Twitch that aim to strike a balance that meets the expectations and needs of all members of our community. Our goal is not to stifle creativity or expression, but we expect everyone on Twitch to hold themselves and their communities accountable.
How can you enforce the sexual harassment policy in Whispers?
Whispers are private communications between users, and as such, we rely on user reports in order to take action on inappropriate behavior being conducted over Whispers. In addition, we offer several tools to help users control how other users contact them. All users can use our “Ignore” feature to stop seeing messages from a specific user. In addition, users can opt to “Block Whispers from Strangers” to prevent any user that they do not have a relationship with from messaging them. Please see our help article for more information.
Now that Twitch has updated the Community Guidelines around Harassment and Hateful Conduct, are there any plans to update how those guidelines apply to conduct off-Twitch?
We are currently working on a broad update to our Off-Service Enforcement policy to provide more clarity about how we define and evaluate user conduct that takes place off-Twitch. We will share more information on the updated policy when it is finalized in 2021.
Does your new policy prohibit the use of terms like “simp,” “incel,” and “virgin” universally? And does this apply to emotes as well?
Our policy prohibits the use of terms like “simp,” “incel” or “virgin” specifically when they are being used to negatively refer to another person’s sexual practices. Using these terms on their own wouldn’t lead to an enforcement but we would take action if they were used repeatedly in a harassing manner.
We deny emotes related to these terms and take them down when they are reported to us. We have a stricter policy on emotes overall because they can be used across Twitch so we take more proactive measures to minimize the potential for harm.