01
Galleries, fixtures coverage and prospectus material now supply identifiable training data for generative tools. Existing image consent regimes were not written with synthetic generation in mind.
02
Indecent synthetic imagery of named pupils can be generated in minutes by a peer, a parent or an outside actor. The school's response framework must assume this is a present, not future, threat.
03
Short audio samples — assembly recordings, podcasts, public lectures — are sufficient to clone a head's voice. Impersonation against parents, staff and finance functions is an emerging vector.
04
Pastoral notes, SEND records, safeguarding minutes and references are routinely pasted into public AI tools. The data leaves the school's control irretrievably.
05
Departments adopt AI tools individually. There is no register, no review, and no defensible record of what processes pupil data. Governance has no line of sight.
06
When an incident occurs, the first hours determine the reputational outcome. Schools without rehearsed protocols default to delay, ambiguity and improvisation.
07
Parents do not assess your technology decisions. They assess the visible composure of the institution. AI readiness is now part of that assessment.