Keeping Children Safe in Education (KCSIE) 2026 introduces substantial changes for schools, colleges and MATs. This year, one of the stronger themes running through the guidance is early risk detection. For SLTs, DSLs, IT and teaching staff, this means being able to quickly identify students at potential risk and intervene before concerns have a chance to escalate.
This article explores 3 specific concerns for which KCSIE requires early risk detection, and outlines the tool that has become essential for education settings to achieve compliance in this area.
“Images may be (...) digitally altered or wholly generated using artificial intelligence, including what are sometimes described as ‘deepfakes’ or ‘deep nudes’.” - Nudes and semi-nudes, including where generated by AI
“All staff should be aware of (...) the nonconsensual making or sharing of nudes or semi-nudes, especially around chat groups.” - para. 23
Deepfakes are images and videos that have been digitally manipulated using AI to present scenes or scenarios that are not real. AI technology can mimic the likeness of a particular person in a highly convincing way, which makes deepfakes extremely harmful when used to create, for example, nude images.
In the UK it is a criminal offence to request or create intimate images of a person without their consent, and this includes AI deepfakes. Despite this, deepfake media targeting both students and staff is a rapidly growing problem in schools and colleges, which is why the risk is now highlighted in KCSIE.
As explained in the guidance, evidence of deepfake activity is often revealed in group chats between peers. Depending on the safeguarding technology settings have in place, a deepfake incident therefore may not become apparent until the content has been shared across multiple devices - by which point the impact on victim wellbeing and school or college reputation has already been felt.
“If a child is struggling with their mental health, self-harming, has an eating disorder or is experiencing suicidal ideation and making plans to end their life, there are likely to be potential warning signs which education staff are well placed to recognise (...) and offer support and vital early intervention.” - para. 48-49
“Education staff can support the fulfilment of 4 key roles: (...) observing pupils and identifying early those who may be experiencing mental health problems or being at risk of developing one, ensuring early targeted support is provided…” - para. 228
KCSIE 2026 further emphasises that mental health concerns should be viewed as potential safeguarding risks. This is particularly true of issues like self-harm, eating disorders and suicidal ideation.
Staff are guided to look for warning signs in how students present or behave that may indicate such mental health concerns. However, this approach alone is unlikely to facilitate the “early targeted support” that is called for. By the time a student shows visible signs of these issues, the risk has escalated to a serious degree. In some cases, students can keep warning signs hidden until it is too late.
This is because today’s students are digital natives who live out many of their highs and lows in digital spaces. A student experiencing desires to self-harm, for example, is now more likely to confide in an AI chatbot or discuss urges with like-minded peers online than show outward distress or confide in an adult.
To spot these risks early and offer effective interventions, schools and colleges need visibility of these online spaces, rather than relying on physical supervision alone.
“Schools and colleges should be aware of the importance of: recognising the escalatory nature of misogyny and the benefits of early identification to support their approach to minimising the risk of HSB, sexual harassment and sexual violence.” - para. 530
“Early intervention in inappropriate behaviour such as misogyny can help to prevent escalation to more serious abuse such as sexual harassment and sexual violence.” - para. 538
Mentions of “misogyny” have more than tripled (from 3 to 11) in KCSIE guidance this year. This is a reflection of just how widespread and harmful misogynistic attitudes and behaviours have become, particularly amongst young men.
While the guidance explicitly calls for “early identification” and “early intervention” to tackle misogyny, without the right training and technology, identifying it is much easier said than done.
Early signs of misogyny can be very subtle, and the language used by groups who promote these attitudes is deliberately coded and constantly changing. It also tends to emerge and flourish in online spaces - through exposure to misogynistic content and influencers.
Without the ability to detect students using misogynistic words or phrases online, education settings will struggle to prevent the spread of misogyny and the harmful behaviour it causes.
Many of the safeguarding risks listed in KCSIE now appear first, and sometimes only, online. While filtering is crucial to block access to harmful websites and pages, filters cannot spot risks in the main spaces where harmful student behaviours are revealed: messaging apps, email, AI chatbots, Word documents, and more.
This is why the Department for Education requires settings to also have monitoring in place. Digital monitoring identifies risks in what students do, say and share online, and alerts the relevant DSL so they can intervene early. As more and more safeguarding risks develop in digital spaces, it has become critical to compliance for schools, colleges and MATs.
Digital monitoring can detect the risks discussed in this article early by: