A ban on school AI use often just moves to a personal phone

The classroom note

A family reading a note home about a new "AI-assisted" homework tool usually wants one thing first: a plain sense of what their child is actually doing with it. A survey of how teenagers already use AI for schoolwork, read alongside department guidance on informing families, as retrieved on 16 September 2026, suggests that conversation is happening less often at school than parents might assume.

What the evidence says

Common Sense Media's survey of more than 1,000 US teens, published 18 August 2026, found 70% report using AI for schoolwork, and among those, 63% say they get an outright answer from AI rather than using it to build understanding. Only 27% say a teacher "has ever talked with them" about safe AI use; 73% say they have not. The same survey found 44% of AI-using teens have had an AI tool blocked on a school network or device, and 59% of that group "just switched to a personal device" to get around it - meaning a school-side block does not, on the survey's own figures, reliably reduce use, only relocate it to a device outside the school's view.

The implementation question

That circumvention finding is precisely why family conversation is a separate mechanism from school policy, not a restatement of it: a ban on a school network cannot reach a personal phone at home. The Australian Framework for Generative AI in Schools names this directly under its transparency principle, stating that parents and carers should have "access to clear and appropriate information" about generative AI, and that school communities should be "appropriately informed" when such tools affect them. The UK Department for Education's guidance adds a narrower, practical warning worth passing to families directly: AI tools can be used to draft "emails from the school to parents that seem realistic," and where a school itself uses AI on pupil data, parents or guardians should "understand that their personal data is being processed."

What holds and what fails

What holds: naming specific, checkable things to tell families - which tools are sanctioned, what happens to their child's data, and why a device block is not a complete answer - does more than a general reassurance that "AI is being handled." What fails: treating a network-level restriction as equivalent to a genuine reduction in student use, when the survey's own respondents describe routing around it. This is an editorial reading of a survey and two guidance documents, not a script for a parents' evening.

• Has a teacher actually discussed safe AI use with our child, or is that assumed to be happening?
• If a tool is blocked on the school network, what happens on a personal device at home?
• What does the school tell families about how a child's data is used inside any AI tool it adopts?

A note home naming these specifics is more useful to a family than a policy summary, precisely because the survey suggests the conversation is not happening as often as the block on the school network implies.