The latest incident in Ceuta sent hate speech on social media soaring, with 9,590 posts detected by the Spanish Observatory on Racism and Xenophobia (OBERAXE) in just two days, coinciding with the mass arrival of migrants on 30 and 31 July, which the government put at around 80,000 people.The Ministry of Inclusion, Social Security and Migration says this spike pushed the number of hate-content items recorded in July to 57,406, up from 40,800 in June.The monthly report notes that people from North Africa remain the main target of hate speech on social media, accounting for 68% of the content detected in July. Messages targeting immigrants made up 19% of the total, followed by those directed at Muslims (16%), people of African descent (10%) and Latin Americans (7%).Public safety fears and the World Cup drove hate speech in JulyPublic safety was the main focus of hate speech in July, accounting for 48% of the content analysed, much of it broadly linking migration with crime.Sport accounted for another 34% of the posts, in a context shaped by the Football World Cup, which fuelled racist, xenophobic and Islamophobic content targeting players, national teams and fans. Migration policies were the focus of 17% of the content, followed by posts about sexual assaults (5%), terrorism (4%) and the arrival of boats (1%).OBERAXE, which reports to the State Secretariat for Migration, notes that four in ten posts dehumanise or seriously degrade the groups targeted. Some 38% contain threats, while 15% call for their expulsion and 4% incite violence. A further 2% seek to discredit them and 1% praise those who attack them.Platforms improve removal of hate contentThe report says that 60% of the content used openly aggressive language, with insults, threats and slurs. Some 34% relied on irony or sarcasm, while 33% used images, videos, memes or coded language, tools that facilitate its spread and make automatic detection harder.The OBERAXE report highlights an improvement in platforms’ removal of hate content, with platforms deleting 12% of user-reported posts, up from 10% in June.Some 9% was taken down within the first 24 hours, while another 3% disappeared over the following week. The role of the “trusted flagger”, which literally means ‘trusted reporter’, enabled the removal of 58% of reported content, seven points higher than the previous month.By platform, X led takedowns with 93%, followed by TikTok (88%), Facebook (74%) and Instagram (71%). YouTube lagged well behind at 14%, although Instagram stood out by more than doubling its removal rate compared with June.