FIFA marked the International Day for Countering Hate Speech on June 18 before a match in Mexico.Paul Childs/Reuters
FIFA says the number of abusive social media posts is on the rise at this World Cup, with 11 per cent of the abuse racially motivated.
Through FIFA’s Social Media Protection Service, six million posts and comments caught the attention of artificial intelligence during the group stage of the tournament with some 225,000 flagged for human review – a 130 per cent increase over the same period at the 2022 edition in Qatar.
Of the 225,000 referred for review, 89,000 were deemed abusive. That marks a 13-fold increase over the 2022 tournament when 6,700 abusive comments were identified.
FIFA says the reviews identified 1,000 accounts for further investigation with more than 100 deemed to “pass the legal thresholds for preparing legal case files against them.”
Racially motivated attacks were the most common form of abuse.
Cathal Kelly: Outmatched all over the field, Canada needs a miracle to beat Morocco
“This is not only a three per cent rise when compared to the equivalent stage of the FIFA World Cup Qatar 2022, but it also marks a significant increase in the objectively worst, most offensive material,” FIFA said in a press release.
There have also been examples of politically motivated abuse, with the Iranian team among those targeted. Abuse related to gambling has also increased.
In all, two million comments and posts – including spam content and abuse generated by bots, or fake accounts – were “hidden,” removed from view to everyone but the sender.
While the increase in abusive posts is no doubt related in part to the expanded 48-team tournament, with 72 group-stage games compared to 64 in the entire tournament four years ago, FIFA says “racial abuse is growing and has become a persistent threat to the well-being of players.”

FIFA says it has removed over 30 million harmful comments directed at players, coaches and officials since launching their project with Signify in 2022.Luke Hales/Getty Images
The ways of delivering abuse over social media have also increased, with more platforms, not to mention weaponized emojis and use of AI-generated videos and images.
FIFA’s social media watchdog has been in operation for four years, starting with the Qatar World Cup.
The world governing body selected Signify, a data science company, to design and run the service. It uses artificial intelligence to monitor the social media accounts of every player, coach and match official at the tournament.
Any content deemed abusive, discriminatory, threatening or violent is referred to Signify’s analysts, who then decide whether further investigation is needed.
Such concerning material is shared with FIFA to determine the next step, which can range from contacting the social media platforms to have the post removed or reporting posts to member associations, to asking law enforcement to get involved.
In Morocco, Canada gets the giant it sought in World Cup knockout play
Different laws in different countries mean that the legal response can vary. Plus, identifying those behind abusive posts can be difficult. But sometimes information gleaned from other posts in the account can lead to identification.
Member associations and players can also ask to have their social media account monitored year-round by the service to have concerning posts hidden.
The moderation service provided for team accounts automatically hid approximately 181,000 “hateful comments” during the group stage of the tournament.
The core protection service only involved social media posts that are public.
But at this tournament, players, coaches and officials can share direct messages with the service to be reviewed. The service can also monitor such messages directly if the individual gives their permission.
The service goes into high gear during events like the World Cup but also operates year-round, monitoring players, coaches and match officials, who are the target of some of the worst abuse.
FIFA says the service has removed more than 30 million abusive posts and comments in more than 50 languages since it was launched.