Perpetrators of violent attacks policy

X · Official agreement

Read the agreement

12 of 17 cards in Content & IP Rights

Content & IP Rights · Context

X says it limits perpetrator content to curb harm and prevent publicizing of messages

X explains that it wants people to be able to express compassion, condemn attacks and discuss their impact. It also acknowledges that some people share manifestos or similar content to express outrage or condemnation of the perpetrator's possible motives. X then states that perpetrators' violence and the rationale in manifestos may endanger targeted people and may inspire future attacks, and that exposure may harm viewers. For these reasons X says it limits dissemination of perpetrator-produced hateful views to keep perpetrators from publicizing their message. This section is background to the rules below, and the rules themselves are set out in the later sections.

Supporting quotations

many want to express compassion for victims, condemn the attacks and/or the perpetrators, and discuss how these incidents impact people and their communities
may further jeopardize the physical safety and well-being of those who are targeted, and has the potential to inspire future attacks
their dissemination should be limited in order to prevent perpetrators from publicizing their message

Qualifications & exceptions

  • Background statement of reasons; the operative rules are in other sections.
Also annotated in the same passage (1)

Content & IP Rights · Restriction or exception

X may remove Posts with manifestos even when not abusive, but may allow some newsworthy content

X says it may remove Posts that include manifestos or similar material produced by perpetrators, even if the context is not abusive. X also says it may allow newsworthy content, but only if the content does not do any of three things: convey suggestions about how to arm oneself and choose targets, share hateful slogans, symbols, memes or hateful conspiracy theories, or outline the perpetrator's ideology, tactical choices or plan of attack. The wording is permissive on both sides ('may remove', 'may allow'), so the overview does not promise either outcome for a particular Post.

Supporting quotations

we may remove Posts that include manifestos or other similar material produced by perpetrators, even if the context is not abusive
we may allow newsworthy content if it does not:
Convey suggestions about how to arm oneself and choose targets;
Share hateful slogans, symbols, memes, and/or hateful conspiracy theories;
Outline the perpetrator’s ideology, tactical choices, and/or plan of attack.

Qualifications & exceptions

  • Newsworthy content may be allowed only if it avoids all three listed things.
  • 'May' on both removal and allowance: X keeps discretion.

Not stated here

  • The overview does not say how X decides whether content is newsworthy.
Also annotated in the same passage (1)

Content & IP Rights · Context

A manifesto is a perpetrator's statement of motivation, views or intent, in any form and at any time

X defines a manifesto as a statement by a perpetrator outlining their motivation, views or intent to engage in a violent attack. It can be a written document, social media post, audio recording, video, external link, letter or other form of content. It may be shared in the aftermath or at any period before an attack. A manifesto can be linked to the event through a statement of warning or intent. The definition is about what the perpetrator says, not about its format or timing.

Supporting quotations

We define a manifesto as a statement by a perpetrator outlining their motivation, views, or intent to engage in a violent attack.
A manifesto can be in the form of a written document, social media post, audio recording, video, external link, or letter or other forms of content.
It may be shared in the aftermath, or at any period before a violent attack.

Qualifications & exceptions

  • Statement must come from a perpetrator and outline motivation, views or intent to engage in a violent attack.
  • Covers statements made before an attack, as well as afterwards.
Also annotated in the same passage (1)

Content & IP Rights · Restriction or exception

Full-length manifestos are prohibited in any context, and excerpts except in a newsworthy context

X lists examples of content it may action: sharing or linking full-length manifestos, regardless of context, and any Posts containing excerpts of a reproduced manifesto (edited or sliced) or of the original manifesto, except when shared in a newsworthy context. The section says violations can occur through Posts, Spaces, images and videos, including live video, and that the list is not limited to the examples. The user-facing sentence says 'you may not Post' these things, while the section header says X 'may remove' and 'may action' them, so X states the prohibition and keeps discretion over enforcement.

Supporting quotations

Additionally, you may not Post any of the following:
We may remove content containing manifestos and other content created by individual perpetrators or their accomplices.
Violations can occur via Posts, Spaces, images, and videos, including live video.
Sharing or linking full-length manifestos, regardless of context
Any Posts containing excerpt(s) of a reproduced manifesto (edited or sliced) or of the original manifesto, except when shared in a newsworthy context.

Qualifications & exceptions

  • Full-length manifestos: regardless of context.
  • Excerpts: not prohibited when shared in a newsworthy context, except for the always-prohibited excerpts in the next card.
  • Applies to Posts, Spaces, images and videos, including live video.
  • Examples are 'not limited to' those listed.

Not stated here

  • The source does not define 'newsworthy context' beyond this wording.

Content & IP Rights · Restriction or exception

Three kinds of manifesto excerpt are always prohibited regardless of context

X states that, regardless of context, it is always prohibited to share excerpts that convey suggestions about how to arm oneself and choose targets, excerpts that share hateful slogans, symbols, memes and/or hateful conspiracy theories such as the Great Replacement Theory, and excerpts outlining the perpetrator's ideology, tactical choices and/or plan of attack. The newsworthy-context exception for ordinary excerpts does not apply to these three kinds. The three kinds match the three things the overview says newsworthy content must not do.

Supporting quotations

Regardless of context, it is always prohibited to share any of the following:
Excerpt(s) that convey suggestions about how to arm oneself and choose targets.
Excerpt(s) that share hateful slogans, symbols, memes, and/or hateful conspiracy theories such as the Great Replacement Theory.
Excerpt(s) outlining the perpetrator’s ideology, tactical choices, and/or plan of attack.

Qualifications & exceptions

  • Applies 'regardless of context'.
  • Limited to excerpts of the three listed kinds.

Not stated here

  • The list formatting does not show clearly whether the media items that follow are meant to sit under the same 'regardless of context' sentence; the media item has its own 'in all instances' wording.

Cross-references

Content & IP Rights · Restriction or exception

Perpetrator-generated media related to the attack is prohibited in all instances

X prohibits sharing perpetrator-generated media related to the attack in all instances. The listed examples, which are not exhaustive, are media filmed by the perpetrator of the attack, memes, stickers or graffiti shared or produced by the perpetrator, and media of the weapons used in the attack. The policy also lists other perpetrator-generated content that is not associated with the violent attack but is specific to the promotion or support of violence.

Supporting quotations

Sharing perpetrator-generated media related to the attack in all instances.
Media filmed by the perpetrator of the attack
Memes, stickers, or graffiti shared and/or produced by the perpetrator
Media of the weapons utilized in the attack
Other perpetrator-generated content that is not associated with the violent attack but that is specific to the promotion or support of violence.

Qualifications & exceptions

  • 'In all instances' for perpetrator-generated media related to the attack.
  • Unrelated perpetrator content is covered only if specific to the promotion or support of violence.
  • Examples are 'not limited to' those listed.

Not stated here

  • The policy does not say whether 'media of the weapons' must come from the perpetrator.

Content & IP Rights · Restriction or exception

Bystander content showing the attack as it takes place is prohibited

The list includes bystander-generated content of the attack as the attack is taking place, such as content that displays a moment of the assault or death, dead bodies, content that identifies victims, or content that depicts the perpetrator(s) conducting the attack. The restriction is tied to content captured while the attack is taking place. The source does not state whether the newsworthy-context exception that applies to manifesto excerpts also applies to this item.

Supporting quotations

Bystander-generated content of the attack as the attack is taking place, such as content that displays a moment of the assault or death, dead bodies, content that identifies victims, or content that depicts the perpetrator(s) conducting the attack.

Qualifications & exceptions

  • Applies to bystander content of the attack 'as the attack is taking place'.
  • The examples are introduced by 'such as', so they are not a closed list.

Not stated here

  • Whether any newsworthy or other exception applies to this item is not stated.

Cross-references

Content & IP Rights · Company practice

X may place a warning screen over some media, including media depicting a perpetrator

X says it may put a sensitive media interstitial over some media. The interstitial puts images and videos behind a warning message that must be acknowledged before the media can be viewed, so people who do not want to see sensitive media can avoid it or decide before viewing. The listed types of media X may interstitial, which are not a closed list, include media depicting a perpetrator and excerpts of the attacker's manifesto that are shared as part of a news report.

Supporting quotations

We may put a sensitive media interstitial over some media.
The interstitial places images and videos behind a warning message that needs to be acknowledged before the media can be viewed.
Media depicting a perpetrator
Excerpts of the attacker’s manifesto that are shared as part of a news report

Qualifications & exceptions

  • 'May' and 'some media': X does not promise a warning on every item of the listed types.
  • Examples are 'not limited to' the two listed.
  • Manifesto excerpts are listed only when shared as part of a news report.

Not stated here

  • The policy does not say how X chooses which media receive an interstitial.

Content & IP Rights · Company practice

X may label URLs that link directly to documents believed to be manifestos

X says it may label URLs that link directly to documents believed to be a manifesto, to prevent those URLs from being shared on X. The trigger is a belief that the linked document is a manifesto, using the policy's own definition of a manifesto, and it applies to URLs that link directly to such documents. The policy does not say what a label looks like or whether a labelled URL can still be posted.

Supporting quotations

We may label URLs linking directly to documents believed to be a manifesto to prevent these URLs from being shared on
We define a manifesto as a statement by a perpetrator outlining their motivation, views, or intent to engage in a violent attack.

Qualifications & exceptions

  • Applies to URLs linking 'directly' to documents 'believed to be' a manifesto.
  • 'May': discretionary.

Not stated here

  • The effect of a label on whether the URL can still be shared is not stated beyond the aim of preventing sharing.
Also annotated in the same passage (1)

Content & IP Rights · Company practice

X may denylist hashtags that identify perpetrators or are used uniquely to surface manifestos

X says it may denylist hashtags identifying perpetrators on trends, to reduce the visibility of perpetrators' identities. It may also denylist all hashtags used uniquely to surface manifestos. The first denylisting is described as applying on trends, so it limits the visibility of those hashtags there, and the source does not say that the hashtags are blocked from Posts.

Supporting quotations

We may denylist hashtags identifying perpetrators on trends to reduce the visibility of perpetrators’ identities, as well as all hashtags that are used uniquely to surface manifestos.

Qualifications & exceptions

  • Hashtags identifying perpetrators: denylisted 'on trends'.
  • Manifesto hashtags: those 'used uniquely' to surface manifestos.
  • 'May': discretionary.

Not stated here

  • What denylisting means beyond trends is not defined.

Content & IP Rights · Restriction or exception

Content depicting listed kinds of violence is not a violation of this policy

X lists content depicting the following as not a violation of this policy: use of force by law enforcement and military personnel, as part of the scope of their official functions, that resulted in casualties; violence against military personnel and law enforcement; human rights abuses; violent attacks that are part of an armed conflict; violence used as a means of self defense; violent attacks where there is probable doubt about the intent behind the attack; and cases of vandalism and attacks resulting in damage to essential infrastructure. This policy does not reach these, but the policy's closing section says X also removes content that violates other rules, such as the Violent Speech policy.

Supporting quotations

Content depicting any of the following:
The use of force by law enforcement and military personnel, as part of the scope of their official functions, that resulted in casualties;
Violence against military personnel and law enforcement;
Human rights abuses
Violent attacks that are part of an armed conflict
Violence used as a means of self defense
Violent attacks, where there is probable doubt about the intent behind the attack
Cases of vandalism and attacks resulting in damage to essential infrastructure

Qualifications & exceptions

  • Applies to content depicting these kinds of violence, under this policy only.
  • Use of force by law enforcement and military personnel counts only 'as part of the scope of their official functions'.
  • Attacks where there is 'probable doubt about the intent'.

Not stated here

  • The policy does not say how these exclusions interact with its separate prohibition on bystander content showing the attack as it takes place (for example content that displays dead bodies).

Cross-references

  • contrast (model-proposed): Bystander content showing the attack as it takes place is prohibited — Content depicting violence against law enforcement or in armed conflict is listed as not a violation, while bystander content of an attack as it takes place, including dead bodies, is listed as prohibited. The source does not say how the two apply to the same content.

Content & IP Rights · Restriction or exception

Sharing a few sentences from a manifesto in a newsworthy or awareness context is not actionable

X lists sharing a few sentences from the manifesto(s), either in a newsworthy context or to bring more awareness to the rise of hate or violence towards a specific group, as not a violation. Its example is that quoting 'the shooter said he wants to eradicate x group' is not actionable under this policy. This sits beside the rule that excerpts are prohibited except in a newsworthy context and the three kinds of excerpt that are always prohibited regardless of context, and the source does not explain how an excerpt that fits this item but also conveys the perpetrator's ideology is treated.

Supporting quotations

Sharing a few sentences from the manifesto(s), either in a newsworthy context or to bring more awareness to the rise of hate or violence towards a specific group.
“the shooter said he wants to eradicate x group” is not actionable under this policy.

Qualifications & exceptions

  • Limited to 'a few sentences'.
  • Must be shared in a newsworthy context or to bring more awareness to the rise of hate or violence towards a specific group.

Not stated here

  • The source does not say how this item applies to excerpts that outline the perpetrator's ideology, which another section says are always prohibited regardless of context.
  • 'A few sentences' is not quantified.

Cross-references

  • contrast (model-proposed): Three kinds of manifesto excerpt are always prohibited regardless of context — One item says sharing a few sentences from a manifesto for newsworthy or awareness purposes is not actionable; another says excerpts outlining the perpetrator's ideology are always prohibited regardless of context. The source does not say how the two apply to the same excerpt.