When we hear the word 'deepfake', many of us might think of celebrities, political misinformation, or the kind of AI 'slop' that now features heavily on our social media feeds. We might be less likely to think about school pupils. Yet emerging research suggests that AI-generated sexual deepfakes are an issue that schools increasingly need to understand and be prepared to address.

AI-facilitated abuse is a growing concern, with reports of deepfake imagery being used maliciously against both pupils and teachers. Whilst anyone can be targeted, research suggests that sexual deepfake abuse is strongly gendered, with girls and women disproportionately affected.

Within UK schools, AI-facilitated image-based sexual abuse most commonly involves groups of boys taking existing photographs of female peers from social media and using generative AI tools to create fake nude images. There have also been reports of criminal gangs targeting schools, threatening to take publicly available images of children from school websites and social media and turn them into sexually explicit images for the purposes of blackmail. In the UK, producing this kind of material is a criminal offence, though globally, laws are not consistent. Whilst deepfake images are not real, the harm they cause is. Emotional, reputational and developmental consequences can be profound and anyone who experiences this form of abuse deserves to be taken seriously and supported.

With her team, our latest Researcher of the Month, Professor Tanya Horeck, has recently published one of the first empirical studies to examine AI-generated sexual deepfakes within UK schools. Using qualitative data from interviews with secondary school staff and pupils at eight UK schools, Professor Horeck explored what young people and school staff understand about sexual deepfakes, how schools are currently responding, teachers' confidence and preparedness and what support schools need. The findings revealed significant gaps in knowledge and understanding, alongside a strong appetite among both students and teachers for more information, guidance and education.

It's important to note that, as an early qualitative study in a rapidly developing area, the research involved only a small number of participants and was designed to explore experiences and perspectives rather than establish how widespread particular views or practices are.

Summary

Professor Horeck and her team found that understanding of AI-generated sexual abuse was limited among both pupils and staff. Young people tended to associate AI primarily with tools such as ChatGPT, which they used for schoolwork, or with obviously fake content encountered on social media. They had much less understanding of generative image technology, how easily it can be used to create sexualised deepfakes or how this might constitute abuse.

The young people interviewed said they had not been explicitly taught about how AI can be used to alter or ‘nudify’ images, or explored the associated questions of consent, ethics and harm. Deepfakes were more likely to be incorporated into broader conversations about AI than addressed specifically as a form of image-based sexual abuse. Teachers also described struggling to keep pace with rapidly developing technology, with some underestimating how accessible and easy to use deepfake tools have become.

Both pupils and teachers connected sexual deepfakes with wider cultures of misogyny. Participants linked the misuse of AI-generated imagery with misogynistic content and influencers encountered online, and the researchers argue that sexual deepfakes should therefore not be understood simply as a technological problem. Instead, they need to be considered within the wider context of technology-facilitated gender-based violence, sexual consent and gendered power imbalances.

Teachers involved in the study were unsure how best to manage incidents in real life. They described uncertainty about how incidents should be categorised and dealt with, including confusion around whether an AI-generated sexual image should be treated in the same way as a genuine non-consensual nude. Responses could focus heavily on the individual incident and supporting a targeted pupil to return to school, rather than recognising what had happened as sexual abuse or considering the wider school culture and preventative work that might be required. The researchers identified a clear need for better policy, staff training and guidance.

There was a strong appetite for better education and support. Rather than suggesting that schools simply need to become experts in every new AI tool, the researchers call for better AI media literacy, embedded within RSHE/PSHE and connected explicitly to consent, privacy, ethics, sexual violence and the law. They advocate a whole-school approach that equips both pupils and staff to recognise AI-generated image-based sexual abuse, understand why it is harmful and know how to respond.

Implications

Tips for schools

Treat deepfake sexual abuse as a safeguarding and sexual violence issue, not simply an AI issue. Professor Horeck advises that schools should explicitly address it as a form of gender-based sexual violence and ensure that it is reflected in relevant safeguarding policies and approaches.

Acknowledge the strongly gendered nature of sexual deepfake abuse. While anyone can be targeted, girls and women are disproportionately affected. Inclusive teaching should recognise that anyone may experience this form of abuse without losing sight of its wider context as a form of gender-based violence. Encourage young people to explore why girls and women are so frequently targeted and connect discussions about sexual deepfakes with wider learning about misogyny and violence against women and girls.

Focus on enduring digital ethics. Schools may understandably feel that they are constantly scrambling to keep up with new technologies, but the fundamental principles young people need remain much more consistent. Build age-appropriate conversations about image ethics, consent, respect, empathy and responsible digital citizenship from primary school onwards. Rather than trying to teach pupils about every new technology or potential harm as it emerges, equip them with an ethical framework that they can apply to new situations.

Challenge the idea that creating sexual deepfakes is a joke or harmless fun. Schools should talk explicitly with young people about the harm that sexual deepfakes can cause and the importance of empathy for those targeted. Given the gendered nature of much deepfake sexual abuse, responses should also challenge dismissive “boys will be boys” narratives and create space to explore the attitudes and behaviours that can underpin this form of abuse.

Embed teaching about sexual deepfakes within wider RSHE work on misogyny and gender-based violence. Help young people to understand sexual deepfakes within the wider context of misogyny and gender-based violence, and address the attitudes and behaviours that can enable these harms rather than treating them simply as an emerging AI problem.

Start conversations early. Age-appropriate discussions about gender stereotypes, gender equality and healthy masculinity can lay important foundations long before young people encounter issues such as sexual deepfakes. Encourage children to question restrictive stereotypes, think about what healthy masculinity can look like and develop respectful attitudes towards people of all genders.

Engage boys as part of the solution, rather than positioning them as the problem. Recognise that some boys may feel blamed or characterised as potential perpetrators by conversations about misogyny and gender-based violence. Create space for open, honest and non-judgemental discussion where boys can question ideas, explore the pressures and expectations associated with masculinity and consider how gender inequality can harm everyone. Helping young people see themselves as having a positive role in creating respectful relationships can make these conversations more constructive.

Consider giving young people some choice over how they discuss sensitive issues. In Professor Horeck’s previous research, pupils were invited to form groups in which they felt comfortable, and this often resulted in girls working with girls and boys with boys. Professor Horeck sees value in allowing that sense of safety, while also creating some opportunities for mixed-gender discussion. Bringing boys and girls together in a youth advisory group, for example, could help young people hear one another’s perspectives and work collaboratively on possible solutions.

Take a whole-school approach. A one-off assembly on deepfakes is unlikely to be enough. Learning should be embedded across the curriculum and developed as pupils get older. Staff also need opportunities to build their understanding of both the technology and the relevant legal and safeguarding issues.

Make young people part of your approach. Professor Horeck strongly advocates listening to pupils rather than relying solely on adults to decide what good digital behaviour should look like. Young people have valuable insights into the digital environments they inhabit and, in her experience, plenty of ideas about how schools might tackle emerging problems effectively.

In the event of an incident, prioritise the victim's/victims' experience. Support and listen to the young person (or people) who has been targeted, rather than judging or interrogating them. Initial conversations should focus on ensuring that they feel heard and supported, and that they have access to further emotional support if they want it.

Avoid questions that could sound blaming, such as asking why someone shared or sent an image. Keep responsibility firmly with the person who created or shared the abusive image. Whether or not a young person has ever chosen to share an intimate image is irrelevant to responsibility for the abuse; creating, manipulating or sharing sexual imagery of someone without their consent is never their fault. With AI-facilitated image-based sexual abuse, the victim may not have shared a sexual image at all – an ordinary, fully clothed photograph can be manipulated without their knowledge or consent.

Look beyond individual incidents to the wider school culture. Responding effectively to a case of deepfake sexual abuse is essential, but Professor Horeck argues that schools also need to consider the wider factors that may enable this behaviour. Explore cultures of misogyny, attitudes towards gender and sexual violence, and young people’s everyday digital practices. Individual incidents can provide an important prompt to consider what broader preventative work may be needed across the school.

Make sure staff understand the law and know how to respond consistently. Deepfake sexual abuse is a serious matter and should always be taken seriously, but Professor Horeck acknowledges that deciding how to respond when children are involved can be complex. In her research, some schools favoured immediate police involvement, while others considered restorative approaches and the wider context of individual cases. Schools should ensure that staff understand the relevant legal and safeguarding framework, have clear procedures to follow and feel confident about escalating incidents appropriately and consistently.

Give staff practical tools to build their confidence. Professor Horeck found that staff can feel apprehensive about discussing sensitive issues such as intimate imagery, particularly when they are unsure which language to use or how to respond to challenging questions or resistance. Providing simple prompts, suggested phrases and practical tips can help staff feel more confident addressing these issues openly with pupils. Find more ideas on the Tooled Up platform.

Never view, download or share the image. Find further guidance on best practice here.

Try using activism as a way into the topic. Professor Horeck found that girls in her research engaged particularly strongly with discussions about celebrity cases of image-based sexual abuse and examples they had encountered through social media and influencers. For older pupils, build on these existing points of interest by inviting them to research campaigners such as Jodie, explore an activist project or investigate how a particular case has contributed to wider debate and change. This can offer a different way into a difficult topic, while showing young people that individuals can raise awareness, influence public debate and contribute to meaningful change. Pupils could even design their own awareness campaign, considering the messages they would want to communicate, who they would hope to reach and what change they would like to achieve.

Teach young people how to be active bystanders. Everyone has a role to play in creating a culture where image-based sexual abuse is challenged rather than ignored or normalised. Try using realistic scenarios involving AI-facilitated abuse to help pupils rehearse what they could do if, for example, they became aware that a sexual deepfake of a peer was being shared.

Give young people agency without placing responsibility for the problem on them. Professor Horeck cautions against implying that young people themselves are responsible for preventing AI-facilitated sexual abuse. Help pupils see the wider media landscape in which these harms occur, including the responsibilities of technology and social media companies. Young people can be part of challenging harmful practices and calling for change, without feeling that tackling the problem rests on their shoulders.

Look out for new resources from Professor Horeck and her team. Professor Horeck is hoping to develop new scenario-based materials specifically focused on active bystander responses to AI-facilitated harm. Schools interested in this work, or in finding out more about the resources she has previously developed, are welcome to contact us and we'll put you in touch.

Tips for parents

Acknowledge that digital intimacy is a normal part of young people’s relationships. Digital intimacy isn't limited to sharing sexual images. It can include messaging a partner late into the night, developing relationships through social media or using technology to communicate feelings and build closeness. For young people who have grown up with digital technology, online and offline relationships are deeply intertwined. Have frank, age-appropriate conversations about what healthy digital intimacy looks like, including consent, respect and how we treat others online.

Ask your teen how AI actually fits into their lives. Rather than approaching AI only through its risks, start with curiosity: How are they using it? What do they value about it? What concerns them? AI chatbots are a useful example. The same rapidly developing technology that raises serious concerns about AI-facilitated sexual abuse may also be experienced by some young people as useful or supportive. Making space for these complexities can open up richer conversations about ethical digital citizenship and help adults understand young people's perspectives rather than assuming we already know what they think. You might find this resource helpful in opening up conversations.

Keep talking. Try to prioritise open communication rather than relying solely on restricting access to technology. Young people may find ways around restrictions, and what matters particularly is that they feel able to come to you when something goes wrong. Stay curious about their digital lives, keep conversations going and aim to create a relationship in which they know they can ask for help without immediately fearing judgement or punishment.

Have a conversation with at least one young person or one group of young people this week about AI and AI facilitated harms. Actively listen to young people because we have a lot to learn from them.