UNESCO Warns AI Is Changing Online Abuse in Schools

AI online abuse in schools is changing how children experience harassment, exploitation and abuse. In a new technical brief, UNESCO warns that online and technology-facilitated violence now affects school safety, learning and wellbeing. The warning reaches beyond cyberbullying. It challenges education systems to improve safeguarding while technology companies and regulators address the digital environments where harm can begin.

For Irish parents and educators, the central question is not whether incidents begin inside a classroom. It is whether children receive effective protection when those incidents follow them into their everyday school lives. UNESCO’s October 2026 guidance calls for coherent policies, properly trained staff, accessible reporting pathways and meaningful cooperation across education, child protection and technology sectors.

Cyberbullying and AI abuse what the evidence shows

UNESCO reports that approximately one in ten learners worldwide has experienced cyberbullying. In a separate survey series covering 44 countries and regions across Europe, Central Asia and North America, reported prevalence increased from 12% in 2018 to 15% in 2022. Those figures describe international study populations and must not be presented as current Irish prevalence.

Chart comparing cyberbullying prevalence of 12 percent in 2018 and 15 percent in 2022 in a 44-country study — Cyberbullying prevalence rose from 12% to 15% in a 44-country survey series. Source: UNESCO (2026), citing HBSC research.

The figures matter because online hostility can continue long after a school day finishes. Threats, abusive group conversations and humiliating posts may reach large audiences within minutes. Moreover, a victim can face renewed harm each time material is shared or displayed to classmates. Schools therefore need a safeguarding response that considers the consequences, not simply the location where an incident started.

AI sexual deepfakes are a serious child protection concern

Research from UNICEF, ECPAT International and INTERPOL estimated that at least 1.2 million children aged 12 to 17 across eleven surveyed countries had experienced manipulation of their images into sexually explicit deepfakes during the previous year. In some of those countries, the estimated prevalence reached one in 25 children. This is not a global total or an estimate for Ireland.

Graphic showing 1.2 million children in eleven surveyed countries affected by sexually explicit deepfake image manipulation — At least 1.2 million children were affected across eleven surveyed countries. UNICEF, ECPAT International and INTERPOL research, cited in UNESCO’s discussion of technology-facilitated harm.

A child does not need to create or share an intimate photograph for this abuse to occur. AI tools can manipulate an ordinary photograph obtained from a social media account or another source. The fabricated image can then be used for humiliation, coercion or extortion. Crucially, a digitally fabricated image does not make the harm imaginary.

Young people may also encounter impersonation, manipulated audio, threatening messages and AI-assisted harassment. However, the availability and documented prevalence of each technique are not identical. Schools should explain the methods without suggesting that every new AI capability is equally common or that every child is experiencing them.

Online violence is not confined to the school gate

UNESCO treats technology facilitated violence as part of the reality of education because its effects extend into participation, relationships and learning. A private messaging group may become a site of humiliation, while a gaming interaction may move to another platform. Likewise, an abusive post created at home may alter how classmates treat one another the following morning.

A sound school response should focus on the child’s safety, preserve relevant evidence where appropriate and follow established child protection procedures. It should not require staff to monitor children’s private accounts or conduct criminal investigations. Serious exploitation, threats or suspected child sexual abuse must be referred to the appropriate safeguarding and law enforcement bodies.

Some children face greater risks and barriers to support

The UNESCO report highlights that harm is not distributed evenly. Girls, learners with disabilities and other marginalised children can experience distinct or more severe forms of technology facilitated violence. Therefore, effective school policies need accessible reporting options, sensitivity to discrimination and support tailored to children’s needs.

One reporting channel will not suit every child. For example, a pupil with a disability may need additional communication support, while another pupil may fear retaliation or exposure if they report sexual harassment. Child centred responses must address these barriers. Furthermore, schools should avoid treating the affected child as responsible for the offender’s behaviour.

Teacher training cannot stop at traditional cyberbullying

Educators increasingly need to recognise AI-generated sexual imagery, impersonation, coercive control, online grooming and repeated harassment across multiple services. Yet awareness alone does not give staff a safe procedure for responding. Training should cover recognising signs of abuse, listening without blame, recording concerns, preserving evidence appropriately and escalating cases.

Teachers also need clear limits to their role. UNESCO calls for workforce support alongside safeguarding policy, rather than simply transferring another responsibility to already stretched schools. In practice, school leaders should know which specialists to contact when technology-enabled behaviour raises child protection or criminal concerns.

AI in education also requires safeguarding checks

AI-powered educational tools promise useful support for learning. Nevertheless, schools must consider how a service handles children’s personal data, content generation, interactions and safety complaints before adopting it. An application marketed for education should not be presumed safe solely because it carries that label.

Procurement and school technology policies should assess age suitability, privacy, reporting mechanisms, moderation, vendor accountability and the risk of harmful outputs. Importantly, these assessments need review as products change. Schools should also consider whether pupils and families understand what information a system collects and how to report something troubling.

Are Irish schools and regulations keeping pace?

Ireland has already strengthened its school bullying framework. The Department of Education and Youth’s Bí Cineálta procedures took effect in September 2025 and explicitly address cyberbullying, sexual harassment and other forms of bullying. Schools must develop policies and involve their communities in prevention and response.

However, recognising cyberbullying in policy does not automatically establish readiness for sexual deepfakes, AI enabled impersonation or the movement of offenders across platforms. Schools should review whether their procedures explain how to respond when manipulated images circulate among pupils. They should also clarify responsibilities, staff training, referrals and communications with affected families.

Meanwhile, Coimisiún na Meán’s Online Safety Code imposes safety requirements on regulated video sharing platform services. Its provisions cover reporting, protection from specified harmful material and relevant age assurance and parental control measures. These obligations are important, but they do not remove the need for effective safeguarding in schools.

In September 2026, the regulator announced an investigation concerning the implementation of age assurance and parental controls on X. That is an investigation into potential compliance issues, not a final finding of wrongdoing. Nevertheless, it illustrates why the practical operation of safety controls matters more than whether a company simply advertises them.

What UNESCO’s recommendations mean in practice

Governments and education authorities

Establish coordinated national policy, assign clear responsibilities and resource school safeguarding systems. They should support staff training, accessible reporting channels and careful measurement of emerging harms. Furthermore, regulators must hold digital services accountable under applicable laws rather than placing the entire burden on schools.

Schools and education providers

Update anti-bullying and safeguarding procedures to recognise AI image abuse and other technology facilitated offences. Staff should know how to receive a disclosure sensitively, record and refer a concern, and support the child afterwards. Schools should also assess the safeguarding implications of educational technology before introducing new tools.

Technology companies

Reduce foreseeable harms through safer product design, usable reporting systems and prompt action against abusive content. Companies must also be transparent about the limits of parental controls and age assurance technologies. Where a platform enables the rapid circulation of abusive content, responsibility cannot be shifted entirely to families.

Parents and carers

Maintain an open interest in children’s online experiences without making them responsible for preventing every threat. Agree on clear boundaries, delay unnecessary social media exposure and encourage children to report troubling interactions early. Parental settings can help, but they do not replace conversations, practical safety education or safer platform design.

Why CODA supports Smartphone Free Childhood

Children of the Digital Age supports the smartphone free childhood movement, reduced recreational screen time and no social media use for national school children. However, children can encounter digital harms through shared devices, gaming, messaging and other people’s actions. Therefore, removing unnecessary exposure and building knowledge must work together.

Online safety education also needs to move beyond generic messages about being kind or not sharing passwords. Children should learn what AI-generated impersonation looks like, why image-based abuse is serious and how to get help. Equally, adults need realistic training about platform behaviour, escalation routes and safeguarding responsibilities.

AI online abuse in schools is a warning Ireland should act on

UNESCO’s guidance is not a new Irish prevalence study, nor does it suggest schools can solve technology-facilitated violence alone. Instead, it offers a framework for coordinating prevention, curriculum, workforce support, reporting and external accountability. Its message is timely because the tools used to harm children can change more quickly than school policies or adult awareness.

The challenge of AI online abuse in schools raises a question for Ireland: whether schools have the support and procedures to recognise AI-enabled abuse before a child experiences prolonged harm. If technology companies, regulators and the education system work in isolation, important gaps will remain. Protecting children requires safer platforms, better informed adults and an effective response when an incident occurs.

References

About Children of the Digital Age

Children of the Digital Age provides online safety education and training for schools, parents and organisations. Our guides support practical conversations about online behaviour and the choices people make. However, education and app settings do not replace safeguarding or professional support.

Send a message

Contact Children of the Digital Age to enquire about online safety education, parent sessions or training for your school or organisation.