🔴 Must Know
England: AI moves inside the safeguarding system tomorrow. Keeping Children Safe in Education 2026 comes into force on 1 September. It is the first edition to name AI-generated content, deepfakes and nudification apps as explicit safeguarding risks rather than subsets of general online harm. The definition of nudes and semi-nudes now covers images “digitally altered, or wholly generated using artificial intelligence.” The 4Cs framework now treats generative AI applications that simulate human interaction as a contact risk. Paragraph 171 prescribes an annual filtering and monitoring review, led by a named SLT member with the DSL and IT support, covering all internet-connected devices, with a formal record kept.
Why it matters: this is statutory guidance for all schools and colleges in England, independent schools included, and British-curriculum international schools in the Gulf, Asia and Europe routinely mirror KCSIE. The practical shift is that a safeguarding concern can now involve an image of a pupil who never took or shared a photograph — something most child-on-child abuse procedures do not currently cover.
EU: the high-risk deadline moved, the literacy duty did not. Regulation (EU) 2026/1744, the Digital Omnibus on AI, entered into force on 27 July and defers Annex III high-risk obligations from 2 August 2026 to 2 December 2027. Annex III covers exactly the school uses that matter most: admissions, grading and evaluation of learning outcomes, placement decisions, and proctoring. But Article 4 (AI literacy) and Article 50 (transparency) were left on their original timeline and apply now. European schools have breathing room on conformity assessment for assessment tools; they have none on the duty to ensure staff and students have a sufficient level of AI literacy.
Subscribe to continue reading
Subscribe to get access to the rest of this post and other subscriber-only content.