Online Harms Affecting Children: What the YJB Evidence Pack June 2026 Shows

About the YJB Online Harms Evidence Pack June 2026

The Youth Justice Board for England and Wales (YJB), an executive non-departmental public body of the Ministry of Justice, published its Online Harms Affecting Children Evidence Pack in June 2026. The pack was authored by Adela Sobrepera and Lamia Hachoumi of the YJB's evidence and insights team, and draws on 104 cited academic studies, government data sources, Ofcom publications, Home Office statistics and practitioner evidence.

The purpose of the pack is to synthesise what is currently known about children's exposure to and involvement in online harms, and to draw out implications for youth justice policy and practice in England and Wales. The Southport Inquiry Phase 1 Report, published in 2026, provides a key part of the policy context - the Inquiry highlighted serious systemic failures to scrutinise and respond to concerning online behaviour. The Online Safety Act 2023 provides the current primary legislative framework for platform regulation.

The pack covers four categories of online harm: cyberbullying, online sexual harms, online radicalisation, and technology-facilitated exploitation. A fifth section examines safeguarding needs and response gaps. A sixth section reviews what works in Youth Justice Service interventions. The report is available at yjresourcehub.uk and via GOV.UK.

Social Media and Violent Content: Scale and Platform Design

The Youth Endowment Fund's 2025 Children, Violence and Vulnerability survey, covering a nationally representative sample of 10,387 children aged 13-17 in England and Wales, provides the most current large-scale evidence on children's online exposure. The survey found that 70% of children in this age group encountered violent content on social media in the 12 months to 2025, a significant increase from 55% in the equivalent 2022 survey. Over four in five children had seen conversations about harming specific groups. 84% of children involved in perpetrating serious real-world violence reported having engaged in online discussions about harming specific groups.

The YJB pack highlights a specific and significant finding on content origin: 25% of children reported that violent content was promoted to them by social media platforms, compared with only 6% who actively searched for that type of content. This finding, drawn from the YEF 2024 social media report, directly implicates platform algorithmic design in the exposure of children to harmful material. The pack identifies five specific social media design features that contribute to harm: content recommendation algorithms, friend suggestion systems, infinite scrolling, push notifications, and popularity metrics including follower counts and likes.

The evidence pack separately addresses the role of generative AI. The National Crime Agency has warned that generative AI is increasingly being used in serious organised crime, particularly child sexual abuse. Online child sexual abuse and exploitation accounted for at least 42% of all recorded CSAE offences in England and Wales in 2024, up from 35% in 2023. The pack notes that Ofcom launched a formal investigation into X (formerly Twitter) under the Online Safety Act in January 2026 following concerns about AI-generated child sexual abuse material being produced and shared via the platform's Grok tool. Ofcom reported in March 2026 that many popular platforms were still failing to properly enforce minimum age policies.

Cyberbullying: Scale, Law and Safeguarding

Cyberbullying is defined in the pack as an intentional and repeated act of aggression using electronic devices, where a power imbalance prevents the victim from defending themselves. It lacks a standalone criminal definition in UK law but may be prosecuted under a range of existing legislation including the Malicious Communications Act 1988, Protection from Harassment Act 1997, Protection of Children Act 1978 and the Online Safety Act 2023.

The Crime Survey for England and Wales 2022/23 found that approximately 19% of children aged 10 to 15 - around 847,000 children - experienced an online bullying behaviour. The Health Behaviour of School-aged Children survey for England 2021/22 found that 23% of girls and 18% of boys reported being victims of cyberbullying. Girls are disproportionately affected. The same survey found that around 12% of children reported cyberbullying others - a 10% increase from 2018. The TES Safeguarding Report 2025 found that 66% of school staff and governors across the UK cited cyberbullying as the most common online safeguarding issue they face.

The pack documents the overlap between cyberbullying and other harmful behaviours. Research from the USA links cyberbullying perpetration to a significantly increased likelihood of weapon carrying one year later - though this association disappears at the two to three year follow-up, potentially reflecting a general decline in weapon carrying and cyberbullying as children age. The emotional and psychological impacts include anxiety, depression, post-traumatic stress symptoms and, in severe cases, suicidal ideation. Behavioural impacts can include substance misuse and aggression. Educational impacts include lower academic performance and school avoidance.

Online Sexual Harms: Intimate Image Abuse and the Law

The YJB pack focuses on intimate image abuse as the online sexual harm with the most direct implications for youth justice. Ofsted's 2021 review of sexual abuse in schools and colleges found sexual harassment, including online, to be widespread, with nearly 90% of girls and 50% of boys reporting that they or their peers had received unwanted explicit pictures or videos.

Youth Justice Board data for the year ending March 2025 shows that 54% of all proven sexual offences committed by children aged 10-17 were related to indecent or extreme pornographic images or videos. Within that category, the largest subcategory was the creation of indecent images of children (45% of image-related offences), followed by sending indecent communications (16%), disclosure or distribution of indecent images (14%), possession of prohibited or extreme pornographic images (12%), possession of indecent images of a child (10%), and voyeurism or upskirting (2%). 97% of all proven sexual offences in the youth justice data were carried out by boys.

The legal framework is set out in the Protection of Children Act 1978, the Sexual Offences Act 2003, and the Online Safety Act 2023, which introduced four new criminal offences relating to the sharing or threatening to share intimate images without consent. Critically, these offences apply to children as well as adults. The pack highlights the tension this creates: children who consensually produce or share intimate images among peers may be technically committing criminal offences. The Home Office introduced Outcome 21 in 2016 to allow police to handle consensual image-sharing among children without formal criminalisation, but the pack notes that HMIP inspections found Outcome 21 was sometimes applied inappropriately to more serious offences including rape.

The pack also addresses the growing risk from AI-generated deepfakes. It notes that deepfake intimate image abuse disproportionately affects women and girls, that the use of deepfakes to produce such content is illegal, and that there are growing concerns about children using accessible generative AI tools to create sexually explicit images of their peers. A Crime and Policing Bill amendment announced at the time of publication would require tech firms to remove non-consensual intimate images within 48 hours or face fines.

Online Radicalisation: Prevent Data and Terrorism Arrests

The number of children arrested for terrorism offences in England and Wales has reached its highest recorded level. In the year ending September 2025, 53 children were arrested under the Terrorism Act 2000 and subsequent legislation - excluding arrests linked to the proscription of Palestine Action in 2025, which significantly inflated overall terrorism arrest figures. Excluding Palestine Action, children accounted for 20% of all terrorism-related arrests (52 of 256). The YJB was notified of 12 children charged with terrorism offences in 2023/24, of whom 92% had no previous cautions or convictions.

Prevent referral data shows children aged 11-15 made up the largest group of referrals in the year ending March 2025, accounting for 36% (3,192 of 8,759 referrals where age was known). This reverses a recent upward trend. Children aged 16-17 were the next largest group at 13% with 1,178 referrals. Children aged 11-15 had the highest referral levels across most concern categories, including 35% of Extremist Right-Wing ideology referrals, 23% of Islamist extremism referrals, and over half of referrals linked to "fascination with extreme violence". Referrals spiked sharply in August 2024 following the Southport attack and again between January and March 2025 during the public trial.

The pack draws on UK Government research suggesting the internet plays an increasingly significant role in radicalisation of 15-25 year olds, operating through three mechanisms: information provision, amplification of group polarisation, and legitimising extreme ideology through echoing. However, it also notes that online-only radicalisation tends to reflect lower intent and operational capability than cases involving offline influences, and that online activity can sometimes provide early warning signs that allow intervention before harm occurs.

A significant section of the pack addresses "the Com" - a decentralised, loosely connected online harm ecosystem identified by the FBI, the NCA and the Global Internet Forum to Counter Terrorism. The Com is described as mostly composed of children and young adults aged 11-25, operating across gaming platforms, livestreams, messaging apps and social networks. The Resolver report cited in the pack uses the term "weaponised loneliness" to describe how children's need for belonging is exploited by these networks. The Crime and Policing Act 2026 introduces new Youth Diversion Orders as an early intervention tool for terrorism-related risks including online radicalisation, expected to come into force from autumn 2026.

The Online Safety Act 2023 and Regulatory Response

The Online Safety Act 2023 is the primary legislative framework for platform regulation in the UK. It places new duties on online platforms to reduce illegal content and protect children from harm. The YJB pack notes that concerns remain around enforcement and the extent to which legislative protections translate into meaningful safety in practice. In March 2026, Ofcom reported that many popular platforms were still failing to properly enforce minimum age policies.

The government announced in June 2026 that it is instructing technology companies to introduce device-level controls to prevent children seeing, sending and receiving explicit content, with legislation to follow if companies fail to act within three months. Plans have also been announced to ban social media access for children under 16 from spring 2027.

The pack notes that the digital landscape is evolving faster than policy can respond. Traditional protective factors such as parental supervision appear less effective online - evidence suggests that general adult supervision does not reduce exposure to harmful online content. The pack recommends safety-by-design and teen-by-default platform measures, early intervention and harm-reduction responses that avoid unnecessary criminalisation, media and gaming literacy programmes, and strength-based interventions that promote belonging and critical thinking.

Assessment tools used by Youth Justice Services are being updated. The Prevention and Diversion Tool, the only mandatory risk assessment that includes questions about children's digital lives, will have those digital life questions incorporated into the AssetPlus assessment framework in summer 2026. The Children's Commissioner is undertaking a research project on children referred to Prevent, with findings expected later in 2026.

What is the YJB Online Harms Affecting Children Evidence Pack?

The Online Harms Affecting Children Evidence Pack is a research synthesis published by the Youth Justice Board for England and Wales in June 2026. It draws on 104 cited primary sources to provide an overview of evidence on cyberbullying, online sexual harms, online radicalisation and technology-facilitated exploitation affecting children in England and Wales, with a focus on implications for youth justice policy and practice. The pack is available at yjresourcehub.uk and via GOV.UK.

How many children experience cyberbullying in the UK?

The Crime Survey for England and Wales 2022/23 found that approximately 19% of children aged 10 to 15 - around 847,000 children - experienced online bullying. Evidence from national surveys across England and Wales suggests 8 to 21% of children aged 9 to 16 experience cyberbullying. Girls are disproportionately affected, with the HBSC survey for England 2021/22 finding 23% of girls compared with 18% of boys reported being victims of cyberbullying.

What does the Online Safety Act 2023 do to protect children?

The Online Safety Act 2023 places duties on online platforms to reduce illegal content and protect children from harm. It introduced four new criminal offences relating to sharing or threatening to share intimate images without consent. Ofcom is the regulator responsible for enforcement. The Act requires platforms to conduct risk assessments for child users and implement age-appropriate safeguards. In January 2026, Ofcom launched a formal investigation into X under the Act following concerns about AI-generated child sexual abuse material. Ofcom reported in March 2026 that many platforms were still failing to enforce minimum age policies.

What is the Prevent programme and how does it relate to online radicalisation?

Prevent is a national programme aimed at preventing individuals from becoming engaged in terrorism through radicalisation. Where a referral progresses and an individual is assessed as being at genuine risk, their case may be considered by a Channel panel for tailored multi-agency support. In the year ending March 2025, 8,759 individuals were referred to Prevent where age was known. Children aged 11-15 made up the largest group at 36% of referrals, accounting for 3,192 cases. Referrals spiked in August 2024 following the Southport attack and again during the trial in early 2025.

What is the UK government doing about social media for children under 16?

The UK government has announced plans to ban social media access for children under 16 from spring 2027. In June 2026, the government instructed technology companies to introduce device-level controls to prevent children seeing, sending and receiving explicit content, with legislation to follow if companies fail to act within three months. A Crime and Policing Bill amendment would require tech firms to remove non-consensual intimate images within 48 hours or face fines. Ofcom is responsible for enforcing the Online Safety Act 2023 requirements on platforms.