AI support with clear labelling and human oversight.

Any AI-generated literacy feedback, recommendations or chat support should be labelled clearly and should not replace human educational or safeguarding judgement.