violence and incitement
Israeli Kfir reservists posed for and shared a photo on social media, showing them humiliating a Palestinian man during a raid in the West Bank. They abducted him from his home, blindfolded and handcuffed him, and wrote “The people of Israel live” on his blindfold.
“According to media reports on November 14, Israel’s Cyber Unit sent Meta and other platforms 9,500 content takedown requests since October 7, 2023, 60 percent of which went to Meta. Platforms are reported to have responded with a 94-percent compliance rate, according to an Israeli official.”
Brown, Deborah, and Rasha Younes. Meta’s Broken Promises: Systemic Censorship of Palestine Content on Instagram and Facebook. Human Rights Watch, 2023.
Policy rationale
We aim to prevent potential offline violence that may be related to content on our platforms. While we understand that people commonly express disdain or disagreement by threatening or calling for violence in non-serious and casual ways, we remove language that incites or facilitates violence and credible threats to public or personal safety. This includes violent speech targeting a person or group of people on the basis of their protected characteristic(s) or immigration status. We remove content, disable accounts and work with law enforcement when we believe that there is a genuine risk of physical harm or direct threats to public safety. We also try to consider the language and context in order to distinguish casual or awareness-raising statements from content that constitutes a credible threat to public or personal safety. In determining whether a threat is credible, we may also consider additional information such as a person’s public visibility and the risks to their physical safety.
In some cases, we see aspirational or conditional threats of violence, including expressions of hope that violence will be committed, directed at terrorists and other violent actors (e.g. “Terrorists deserve to be killed”, “I hope they kill the terrorists”). We deem those non-credible, absent specific evidence to the contrary.
We remove threats of violence against various targets. Threats of violence are statements or visuals representing an intention, aspiration, or call for violence against a target, and threats can be expressed in various types of statements, such as statements of
We do not prohibit threats when shared in awareness-raising or condemning context, when less severe threats are made in the context of contact sports, or certain threats against violent actors, such as terrorist groups.
[Universal protections for everyone]
Everyone is protected from the following threats:
- Threats of violence that could lead to death (or other forms of high-severity violence)
- Threats of violence that could lead to serious injury (mid-severity violence). We remove such threats against public figures and groups not based on protected characteristics when credible, and we remove them against any other targets (including groups based on protected characteristics) regardless of credibility
- Admissions to high-severity or mid-severity violence (in written or verbal form, or visually depicted by the perpetrator or an associate), except when shared in a context of redemption, self-defence, contact sports (mid-severity or less), or when committed by law enforcement, military or state security personnel
- Threats or depictions of kidnappings or abductions, unless it is clear that the content is being shared by a victim or their family as a plea for help, or shared for informational, condemnation or awareness-raising purposes
In addition to the universal protections for everyone, all private adults (when self-reported), children and persons or groups of people targeted on the basis of their protected characteristic(s), are protected from threats of low-severity violence.
[Other violence]
In addition to all of the protections listed above, we remove the following:
- Content that asks for, offers or admits to offering services of high-severity violence (for example, hitmen, mercenaries, assassins, female genital mutilation) or advocates for the use of these services
- Instructions on how to make or use weapons where there is language explicitly stating the goal to seriously injure or kill people, or imagery that shows or simulates the end result, unless with context that the content is for a non-violent purpose such as educational self-defence (for example, combat training, martial arts) and military training
- Instructions on how to make or use explosives, unless with context that the content is for a non-violent purpose such as recreational uses (for example, fireworks and commercial video games, fishing)
- Threats to take up weapons or to bring weapons to a location or forcibly enter a location (including but not limited to places of worship, educational facilities, polling places or locations used to count votes or administer an election), or locations where there are temporary signals of a heightened risk of violence.
- Threats of violence related to voting, voter registration, or the administration or outcome of an election, even if there is no target.
- Glorification of gender-based violence that is either intimate partner violence or honour-based violence
We remove:
- Threats against law enforcement officers or election officials, regardless of their public figure status or credibility of the threat.
- Coded statements where the method of violence is not clearly articulated, but the threat is veiled or implicit, as shown by the combination of both a threat signal and contextual signal from the list below.
- Threat: a coded statement that is one of the following:
- Shared in a retaliatory context (e.g. expressions of desire to engage in violence against others in response to a grievance or threat that may be real, perceived or anticipated)
- References to historical or fictional incidents of violence (e.g. content that threatens others by referring to known historical incidents of violence that have been committed throughout history or in fictional settings)
- Acts as a threatening call to action (e.g. content inviting or encouraging others to carry out violent acts or to join in carrying out the violent acts)
- Indicates knowledge of or shares sensitive information that could expose others to violence (e.g. content that either makes note of or implies awareness of personal information that might make a threat of violence more credible. This includes implying knowledge of a person’s residential address, their place of employment or education, daily commute routes or current location)
- Context
- Local context or expertise confirms that the statement in question could lead to imminent violence.
- The target of the content or an authorised representative reports the content to us.
- The target is a child.
POLICY PARADOX
Meta’s Violence and Incitement Standard claims to remove content that “promotes or glorifies violence.” Yet here we see Israeli soldiers proudly documenting war crimes, knees on necks, blindfolded detainees, rifles aimed at children, all permitted under Meta’s watch. These images remain untouched, allowed to circulate freely under Meta’s watch. Meanwhile, Palestinian journalists documenting these same acts are flagged, restricted, or removed entirely within minutes.
- Implicit threats to bring armaments to locations, including, but not limited to, places of worship, educational facilities, polling places or locations used to count votes or administer an election (or encouraging others to do the same) or locations where there are temporary signals of a heightened risk of violence.
- Claims or speculation about election-related corruption, irregularities or bias when combined with a signal that content is threatening violence (e.g. threats to take up or bring a weapon, visual depictions of a weapon, references to arson, theft, vandalism), including:
- Targeting individual(s)
- Targeting a specific location (state or smaller) where the target is not explicit
ERASED EVIDENCE
When Palestinian blood appears on camera, Meta’s systems see “terrorist content.” When Zionist guns appear, they see “national defense.” By silencing testimonies of occupation while allowing Israeli narratives to dominate, Meta shapes the historical record to favor the occupier.
This isn’t content moderation; it’s an algorithmic endorsement of oppression. Each flagged post, each removed account, contributes to a curated narrative where resistance is criminalized, and state violence is normalized. They’re curating genocide.
Meta is complicit in genocide.
- References to election-related gatherings or events when combined with a signal that content is threatening violence (e.g. threats to take up or bring a weapon, visual depictions of a weapon, references to arson, theft, vandalism).
- Threats of high- or mid-severity violence in the defence of self or another human when the criteria below are met.
- Against a person (excluding persons identifiable by name or face, people targeted based on their protected characteristics, and children)
- In the context of home entry or interpersonal violence that is proportional to the violence responded to and is an immediate threat
The potential impact on voice outweighs the risk of imminent violence