X fails to act on 97% of calls for violence against migrants or Muslims
A year after the UK summer riots, Elon Musk's X continues to be a breeding ground for violent anti-Muslim and anti-migrant hate.
About
The social media platform X is failing to act on 97% of posts calling for violence against migrants or Muslims that are reported to the platform, according to new research from the Center for Countering Digital Hate.
The research follows a new report by the Center, Fuelling Hate, that finds X is hosting thousands of calls for violence against Muslims or migrants in reply to well-known hateful influencers one year on from the UK summer riots.
Under X’s rules “explicitly threatening, inciting, glorifying, or expressing desire for violence” is prohibited and violating content is subject to removal or reduced visibility.[1]
To test X’s enforcement of its rules, researchers selected a random sample of 100 posts which call for violence against Muslims or migrants. They partnered with staff working for the Southport MP Patrick Hurley who reported these posts using X’s own reporting tools under the “Violent Speech” category.
After giving the platform one week to respond, each of the posts were checked to see if action had been taken on the post. Researchers found that X failed to act on 97% of the posts, with just three posts displaying a box stating “Visibility limited: this Post may violate X’s rules against Violent Speech.”
| Action from X | One week after reporting |
| No Action | 97% |
| Visibility Limited | 3% |
| Post Removed | 0% |
X enforced its own rules inconsistently. While the platform did limit the visibility of some posts, X failed to take visible action against the majority of posts. Examples of calls for violence against Muslims and migrants that X failed to act on include:
- “I couldn’t hate them more if I tried. When we rise up, the whites, sihks, Eastern Europeans and destroy Islam in the uk. I pray for the day.”
- “You can’t send them home alive, they will come back armed… They must be neutralized either by taking their hands or sending them to live with Alluah.”
- “Death to Islam”
- “Deport or exterminate”
This failure to act highlights significant gaps in X’s moderation of violent hate content, despite clear policy prohibitions against threats and incitement.
How violent posts were reported to X and action measured against them
Researchers took a random sample of violent posts that were identified through an AI-assisted system that was built to flag posts promoting violence towards Muslims or migrants.
Researchers reviewed the sample and removed posts that were incorrectly identified as violent or that were no longer visible on X. Two researchers checked each post to confirm if they contained calls for violence targeted at Muslims or migrants, with any disagreements between researchers discussed to confirm a consensus.
This list of 100 posts was shared with the staff of Southport MP Patrick Hurley, who reported them between 7th July and 8th July 2025 using X’s own reporting tool. All were reported under the “Violent Speech” category, in one of three subcategories, “Incitement of Violence”, “Wish of Harm”, or “Violent Threats” depending on the content of the violent reply.
On 15th July, researchers reviewed each comment to check if X had acted to remove the comment or had reduced visibility. All posts were still accessible, and three of the posts had an added box stating: “Visibility limited: this Post may violate X’s rules against Violent Speech.”
CCDH identified over 4,000 posts promoting violence toward Muslims or migrants
CCDH’s recent report “Fuelling Hate” found that in the year after the 2024 UK summer disorder, calls for violence on X still spread unchecked. CCDH examined the persistence of posts promoting violence on the social media platform X, focussing on violence directed at Muslims and migrants, the main targets of the 2024 disorder.
Following the Southport attack, CCDH research identified six X accounts that were influential in spreading hate and misleading claims during the disorder that followed: Tommy Robinson, Paul Golding, Ashlea Simon, Andrew Tate, Laurence Fox, and Calvin Robinson.
Posts about Muslims or migrants by these accounts continue to act as magnets for violent responses, with CCDH identifying 4,379 violent replies calling for shooting, maiming, or killing of Muslims and migrants.
Researchers used an AI-based system to identify violent replies across three categories: direct incitement to violence, explicit threats, and expressions of a desire for violence. Thousands of calls to violence in these categories remain on X despite being clearly prohibited by the platform’s policies.
In this supplemental study, Southport MP Patrick Hurley’s office reported a random sample of 100 violent replies to the platform. This new study finds that X failed to act on 97% of these violent posts after being alerted to them.
Read CCDH’s full report here.
[1] “Violent Content”, X Help Center, February 2025, https://help.x.com/en/rules-and-policies/violent-content