When digital humiliation follows children beyond the classroom.
Madrid, Spain.
Artificial intelligence is becoming an increasingly prominent instrument of cyberbullying in Spanish schools, appearing in 23.5% of reported incidents identified by students during the 2025–2026 academic year, compared with 14.2% the previous year. The findings, published by the Mutua Madrileña Foundation and the ANAR Foundation, reveal how generative technologies are expanding the methods available to students who harass their peers. Manipulated photographs, fabricated videos and synthetic voices can now turn ordinary personal content into instruments of humiliation. The evidence points to a growing technological dimension of school harassment, although it does not establish that AI caused the overall increase in bullying.
The eighth national report on school bullying examined 18,245 valid student questionnaires and responses from 412 teachers across 236 Spanish educational institutions. According to the findings, 14.1% of surveyed students reported that they or a classmate had experienced bullying, cyberbullying or both, up from 12.3% in the preceding academic year. Among cyberbullying incidents involving AI, manipulated photographs used to produce fabricated videos accounted for 79.5% of cases. Voice manipulation appeared in 57.5%, while identity impersonation was identified in 47.9%.
Girls represented 60.3% of victims in incidents involving artificial intelligence, revealing an important gender dimension that requires further investigation. The psychological consequences of manipulated content can extend beyond the original act of aggression, particularly when fabricated material is circulated among classmates or presented as authentic. Digital content can persist after an incident is reported, creating additional difficulties for victims seeking to recover their privacy and social relationships. These characteristics make AI-assisted harassment particularly challenging for schools accustomed to addressing aggression primarily within physical classrooms.
Social media platforms have become important distribution channels for abusive material. WhatsApp appeared in 78.4% of the cyberbullying cases examined, followed by TikTok at 65.6% and Instagram at 50.4%. The circulation of false information and rumors was the most frequently identified behavior, affecting 71% of cases, while 22.8% of cyberbullying victims reported experiencing harassment for more than a year. These findings illustrate how digital aggression can become prolonged and difficult to contain.
The institutional response faces significant operational constraints. Nearly 85% of surveyed teachers identified insufficient resources as an obstacle to intervention, while 69% cited bureaucratic difficulties or unclear procedures. A further 46.9% reported shortcomings in teacher training. These limitations suggest that introducing new technologies into educational environments without corresponding safeguarding capabilities may widen the distance between emerging risks and institutional responses.
UNICEF recommends preserving evidence, reporting abusive content, informing educational authorities and seeking professional support when harassment affects a student’s emotional well-being. Such interventions become especially important when manipulated images or voices make it difficult to establish authenticity. Schools also need clear procedures for distinguishing fabricated material from genuine evidence while avoiding unnecessary redistribution of harmful content.
Spain’s findings expose a broader challenge for human-centered artificial intelligence. The issue is not simply whether students can access powerful generative tools, but whether educational institutions possess the psychological, technological and organizational capacity to respond when those tools are misused. Effective prevention requires digital literacy, accessible reporting systems, family participation and consistent institutional accountability.
Artificial intelligence has not invented school bullying. It has introduced new capabilities into an existing problem, making the protection of children’s dignity and psychological safety an increasingly complex responsibility.
Phoenix24: clarity in the grey zone. / Phoenix24: claridad en la zona gris.