FIFA Social Media Protection Service
Tackling online abuse at the FIFA World Cup Qatar 2022™
With an independent report finding more than half of the players who took part in the final games of UEFA EURO 2020 and CAF AFCON 2021 suffered some form of online abuse, it was crucial that the Social Media Protection Service was introduced ahead of FIFA World Cup Qatar 2022™. By proactively reporting and moderating abusive messages on behalf of participating teams and players, they – and their followers – were able to focus on playing their part in the greatest show on earth.
Social Media Protection Service at FIFA World Cup Qatar 2022™
Posts / comments analysed
Comments hidden
Posts flagged by AI and reviewed by humans
Posts / comments verified as abusive and reported to platforms
Unique accounts detected sending abusive messages
accounts identified for real-world action
Levels of abuse
Across the FIFA World Cup Qatar 2022™, protective coverage was set up to incorporate the following participants:
864 players and coaches with 2,000 active accounts
129 officials with 31 active accounts
32 teams/associations with 126 active accounts
Coverage spanned across the five major social media platforms of Facebook, Instagram, TikTok, Twitter and YouTube:
Accounts coverage
Platform | Active accounts |
43% | |
26% | |
24% | |
Tiktok | 6% |
YouTube | 1% |
Participants were protected from abusive content in their respective nations’ official language(s), as well as in FIFA’s official languages.
The service’s proactive monitoring capability scanned over 20 million posts and comments on Twitter, Instagram, Facebook, TikTok and YouTube throughout the tournament.
The system flagged 433,696 posts/comments as containing language that had or could be used in an offensive way. This dataset was then reviewed by a team of expert analysts with a double human triage process to ensure that no posts were incorrectly attributed, removing the presence of false positives. This led to a final set of 19,636 posts/comments being confirmed by the service provider as abusive, discriminatory or threatening, which were then reported to platforms for breaking their community guidelines. In many cases, the offending posts were removed by the social platform in question as a direct result of being reported by FIFA.
As seen at UEFA EURO 2020 and CAF African Cup of Nations 2021, on-pitch events and results were a trigger for abuse aimed at participating teams and players, with knockout matches in particular creating some key flash points as teams secured qualification or faced elimination. Players’ political allegiances or personal circumstances were often identified as catalysts for inciting vitriol against them.
While 26% of the total detected abuse was classified as ‘general abuse’, a further 17% was sexual in nature. Sexism (13%), homophobia (12%) and racism (11%) were also particularly prolific.
Violence and threat became more extreme as the tournament progressed with players’ families increasingly referenced and many threatened if they returned to a particular country – either the nation they represent or where they play their club football.
Abuse Categorisation
Detected abusive messages by category: All platforms | % |
General abuse | 26.24% |
Sexual | 17.09% |
Sexism | 13.47% |
Homophobia | 12.16% |
Racism | 10.70% |
Family | 6.56% |
Violence | 5.62% |
Ableism | 2.07% |
Islamophobia | 1.94% |
Dogwhistle | 1.74% |
Anti-GRT | 0.37% |
Transphobia | 0.28% |
Special Terms | 0.26% |
Political | 0.24% |
Antisemitism | 0.18% |
Intracommunity slurs | 0.09% |
Sectarianism | 0.07% |
Content was subject to moderation, monitoring or reporting where it included a reference – whether express or implied – to any one or more of the reasons listed by Article 4 of the FIFA Statutes, where the context may be reasonably concluded to be harmful.
DISCRIMINATION of any kind against a country, private person or group of people on account of race, ethnic, national or social origin, gender, disability, religion, political opinion or wealth, birth or sexual orientation is STRICTLY PROHIBITED.
Additionally, any content which may be deemed to include threat of harm to the subject or their family members was automatically included for assessment.
Discrimination is a criminal act. With the help of this tool, we are identifying the perpetrators and we are reporting them to the authorities so that they are punished for their actions. Our position is clear: we say no to discrimination.
Trends and insights
There were several flashpoints during the tournament, with the largest single of targeted abuse coming following the quarter-final match between England and France. The second-largest volume of abuse was around the final match of the tournament, while the game between Germany and Japan led to the largest spike of abuse during the group stage.
Abuse timeline
Match | Volume | Date |
Qatar v Ecuador | 6,262 | 20/11/2022 |
Mexico v Poland | 6,861 | 22/11/2022 |
Germany v Japan | 8,785 | 23/11/2022 |
Portugal v Uruguay | 4,807 | 28/11/2022 |
Saudi Arabia v Mexico | 8,292 | 30/11/2022 |
Costa Rica v Germany | 8,652 | 01/12/2022 |
Ghana v Uruguay | 7,177 | 02/12/2022 |
Morocco v Spain | 4,206 | 06/12/2022 |
Morocco v Portugal | 10,757 | 10/12/2022 |
England v France | 12,823 | 10/12/2022 |
Post England v France | 7,895 | 11/12/2022 |
Semi Final 1 | 7,215 | 13/12/2022 |
Semi Final 2 | 6,878 | 14/12/2022 |
Final | 11,709 | 18/12/2022 |
On the day of the FIFA World Cup Qatar 2022™ final and the 24-hour period following it, the Social Media Protection Service scanned and analysed more than 1.6 million posts and comments. Of these, more than 45,000 were flagged for human review and 1,239 were reported directly to social media platforms for further action.
Compared to the finals of UEFA EURO 2020 and CAF African Cup of Nations 2021, the FIFA World Cup Qatar 2022™ final was notable for the range of abuse types targeting players online. The finals of EURO 2020 and AFCON 2021, by contrast, were targeted most heavily by racist and homophobic content with 78% of all detected abuse falling into one of those two categories.
Final v. Final abuse categorisation
AFCON 2022 / Euro 2020 Finals | % | FIFA World Cup Qatar 2022™ Final | % |
Homophobia | 40% | General Abuse | 24% |
Racism | 38% | Sexual | 18% |
Islamophobia | 6% | Racism | 18% |
Ableism | 6% | Sexism | 13% |
Other | 4% | Homophobia | 9% |
Xenophobia | 3% | Family | 8% |
Violence | 4% | ||
Ableism | 2% | ||
Transphobia | 1% | ||
Xenophobia | 1% | ||
Dogwhistle | 1% | ||
Anti-GRT | 1% |
Racist and homophobic abuse is typically the most egregious and more easily identifiable/actionable by platforms, whereas more nuanced abuse (such as targeting of family members, more subtle ‘dog whistles’ or abuse sent in non-English languages) makes traditional detection more complex.
France’s players, who finished the tournament as runners-up, suffered the most collective online abuse across the competition. Player and team accounts belonging to Brazil and England were also heavily-targeted by abusers. During the group stage, Mexico had almost twice as many abusive messages targeting their players compared to other teams. Detected abuse in the early stages of the tournament was more issue-based, with England particularly targeted for off-field events, while Germany and the USA also saw high levels of homophobic messaging connected to off-field issues.
As well as being targeted for performance issues and player profile, there was also a noticeable correlation between frequency of content being posted on official channels and volumes of abuse.
Targeted teams
Abuse Categories | Germany | Morocco | USA | Portugal | Uruguay | Argentina | Mexico | England | Brazil | France |
Ableism | 8 | 7 | 58 | 42 | 15 | 32 | 31 | 136 | 59 | 75 |
Anti-GRT | 0 | 0 | 0 | 5 | 0 | 7 | 13 | 4 | 0 | 7 |
Antisemitism | 2 | 0 | 4 | 3 | 0 | 3 | 2 | 19 | 1 | 11 |
Dogwhistle | 8 | 12 | 84 | 14 | 7 | 7 | 9 | 141 | 44 | 58 |
Family | 29 | 51 | 67 | 57 | 189 | 121 | 53 | 163 | 89 | 463 |
General abuse | 101 | 131 | 243 | 342 | 283 | 344 | 697 | 1171 | 903 | 1328 |
Homophobia | 148 | 146 | 146 | 51 | 40 | 77 | 194 | 319 | 121 | 697 |
Intracommunity slurs | 0 | 3 | 0 | 0 | 0 | 0 | 0 | 2 | 0 | 3 |
Islamophobia | 90 | 21 | 13 | 6 | 1 | 1 | 0 | 5 | 5 | 54 |
Political | 6 | 6 | 0 | 2 | 1 | 6 | 0 | 0 | 15 | 27 |
Racism | 44 | 117 | 66 | 60 | 52 | 130 | 103 | 401 | 215 | 762 |
Sectarianism | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 8 | 3 | 2 |
Sexism | 63 | 115 | 69 | 173 | 292 | 264 | 329 | 125 | 724 | 913 |
Sexual | 62 | 159 | 176 | 210 | 289 | 287 | 226 | 284 | 814 | 941 |
Special terms | 3 | 2 | 0 | 0 | 0 | 0 | 0 | 0 | 53 | 4 |
Transphobia | 0 | 6 | 0 | 0 | 0 | 0 | 1 | 1 | 7 | 57 |
Violence | 14 | 28 | 114 | 121 | 42 | 68 | 138 | 241 | 72 | 209 |
Xenophobia | 19 | 6 | 6 | 4 | 0 | 0 | 0 | 23 | 3 | 31 |
Totals: | 597 | 810 | 1,046 | 1,090 | 1,211 | 1,347 | 1,797 | 3,043 | 3,128 | 5,642 |
Football has a responsibility to protect the players and other affected groups around their workspace. Therefore, FIFPRO and FIFA will continue their collaboration but we cannot do this alone – we need all stakeholders to play their part.
Real-world action
As part of the Social Media Protection Service, teams and players participating in the tournament were offered access to software that – with the account holder’s permission – could automatically and instantly hide abusive and offensive comments on Facebook, Instagram and YouTube only.
Through the moderation element of the service, a total of 286,895 comments were hidden from public view before the targeted recipient or their followers had to see their contents.
In order to support real-world action being taken against those who sent online abuse to teams and players taking part in the tournament, Social Media Protection Service implemented a tiering system to grade the severity of the comment/post in question:
Tier 3:
Content breaches platform’s terms of service or community guidance and should be removed
Tier 2:
Member Associations or domestic clubs may wish to act via education initiatives or ticket sanctions
Tier 1:
Jurisdictional law enforcement action may be appropriate
12,618 accounts were recorded sending abuse or threat across the tournament. 1,189 of those accounts were graded as having posted comments that fell into tier one and, of those, 306 have had their identity verified by the service provider.
Of the 12,618 accounts that sent abuse during the tournament, it was possible to identify locations for 7,204. 38% came from Europe (including 17% from the United Kingdom) with the same amount coming from South America and Central America.
Abuse locations
Location | % |
Europe | 38% |
South America | 36% |
Asia | 10% |
Africa | 8% |
North & Central America | 8% |
Verified information about the aforementioned 306 accounts has been made available by FIFA for relevant Member Associations and jurisdictional authorities to support real-world action being taken against those who sent abusive, discriminatory and threatening comments to participating teams and players during the FIFA World Cup.
Click here to read the full report.