Safeguarding: AI in Schools
As artificial intelligence (AI) embeds into modern society, the educational conversation has largely focused on workload reduction for teachers and hyper-personalised learning for students. However, beneath the surface of automated grading and smart tutoring lies a complex network of child safety concerns.
Integrating AI into schools completely alters the landscape of safeguarding. Pupil safety in the era of AI is no longer just about content filters and secure school systems; it requires navigating data exploitation, predictive bias, and entirely new vectors of digital harm.
As evident in the revised statutory guidance Keeping Children Safe in Schools (KCSIE), the fast pace of AI integration and the proliferation of synthetic content and deepfakes present a myriad of new safeguarding themes which all staff need to be aware of.
For example, in cases of child-on-child abuse, students no longer need technical expertise to create highly realistic synthetic media. There is a rise in the creation of non-consensual AI-generated imagery. Inexpensive, easily accessible ‘nudification’ apps and deepfake software allow pupils to manipulate the likeness of their peers or teachers, which can lead to severe bullying and devastating psychological harm.
Outside the school gates, adult offenders are utilising AI as a means of coercion and grooming by generating exploitative material or creating highly convincing, automated personas to groom children on social platforms. As noted by the Internet Watch Foundation (IWF), the sheer velocity and severity of AI-generated threat vectors complicate traditional web-filtering protocols.
As Large Language Models (LLMs) become integrated into educational platforms as ‘24/7 AI Tutors’ or ‘study buddies’, pupils are increasingly interacting with conversational agents that mimic human relationships. While these tools may offer immediate academic assistance, they lack genuine emotional intelligence, a phenomenon researchers call the ‘empathy gap’.
When children anthropomorphise AI, treating a chatbot as an empathetic confidant, they become uniquely vulnerable. AI chatbots do not possess a moral compass or human intuition, presenting significant risks to children, particularly vulnerable children, using them.
AI introduces a fast-paced, rapidly developing online environment which schools must proactively accommodate now. In addition to policy and the development of an agreed internal AI usage framework, the curriculum must pivot to expand the teaching of digital resilience and critical media literacy in order to prepare pupils to handle the emotional and legal realities of deepfakes and algorithmic manipulation. Proactive AI resilience is key and, within this planning, safeguarding must be a key priority for both pupils and staff.
The SSS Learning course AI - The Safeguarding Implications for Schools will be available in September 2026.
SSS Learning
8 July 2026