AI for Parents: Seven Conversations to Have Before Your Child Uses It for Schoolwork

Many parents are being asked to supervise a technology they did not use as students. That creates two easy extremes: ban it completely or assume every use is productive because AI is now everywhere. Neither helps a teenager build judgment.

1. What job are you asking AI to do?

Brainstorming examples, explaining a difficult concept and writing an entire submission are not the same use. Ask the child to name the role the tool played.

2. How will you check whether it is right?

A fluent answer can still be wrong. Build a habit of verification with textbooks, teachers, primary sources or trusted references.

3. What information are you giving it?

UNICEF and UNESCO both emphasise privacy and child-centred safeguards. Students should not casually paste personal, school, health or other sensitive information into tools they do not understand.

4. What part of the work must still be yours?

If the purpose of an assignment is to practise writing, reasoning or problem solving, handing the whole task to AI defeats the learning objective even if the final answer looks polished.

5. Can you explain the answer without the screen?

A simple test of learning is whether the student can restate the idea, defend the choices and answer a follow-up question.

6. What might the system be missing or biasing?

Students should learn that AI systems are built from data and design choices. Outputs can reflect gaps, bias and overconfidence.

7. Does your teacher or school have a rule for this?

AI expectations should be transparent. If a teacher requires disclosure or prohibits a tool for a particular assessment, that matters.

"The parenting goal is not to supervise every prompt. It is to raise a student who knows when to use a tool, when to question it and when to stop."

A family AI agreement can fit on one page

Parents do not need to monitor every chat. Agree on a few durable rules: do not upload private information; do not submit AI output you cannot explain; check factual claims; disclose material assistance when the teacher expects it; and keep some learning tasks completely tool-free when the purpose is practice. The agreement should change as the child and school expectations change.

One useful habit is the 'show your process' conversation. Instead of asking only to see the final homework, ask the child to show one prompt, one answer they rejected or corrected, and the final version. That makes judgment visible without turning the parent into a surveillance officer.

Watch for dependence, not simply frequency

Frequent AI use is not automatically harmful and occasional use is not automatically good. The concern is whether the student is losing the ability or willingness to start a task, tolerate confusion, verify an answer or produce an explanation independently. Those are learning behaviours worth protecting.

• Can the student attempt the first step before asking AI?
• Can they identify a wrong or weak answer?
• Can they explain what they changed and why?
• Can they complete some comparable work without the tool?

Use the teacher's learning objective as the deciding rule

The same AI use can be reasonable in one assignment and counterproductive in another. If the objective is to practise grammar, automatic rewriting may remove the practice. If the objective is to compare arguments, generating candidate arguments for critique may be useful. Parents can ask one powerful question: 'What was this assignment trying to help you learn?'

That keeps the family discussion educational rather than moralistic. The question is not whether AI is good or bad; it is whether the use supports the intended learning and follows the teacher's rules.