The AI they are already using

Not “should my child use AI”. They are using it — for homework, out of curiosity, and in some cases to talk to. The useful questions are which kind, for what, and whether it has started standing in for people.

The posture Banning it outright is the one approach that reliably fails. It drives the use underground, removes you from the conversation, and guarantees they will not ask you when something goes wrong.

Know which kind you are looking at

These get lumped together and they are not remotely the same risk.

• Assistants ChatGPT, Gemini, Copilot and the like. Mostly a homework and curiosity question. Real but ordinary risks: wrong answers delivered confidently, and academic honesty.
• Companion and character apps Where the AI plays a friend, a partner or a character. Built for engagement and emotional attachment. This is the category to actually look at.

Worth knowing the names of what is on the phone, and worth asking directly rather than guessing: is there one of these you talk to, as opposed to ask things?

The demonstration that beats a lecture

Sit down together and ask a chatbot something you know cold — a detail of family history, a local fact, the specifics of something from your work. Ask it for a source. Watch it produce a confident, fluent, wrong answer, and then watch it apologize and produce a different one when challenged.

Five minutes of that does more than any amount of explaining, because the point lands as something your child observed rather than something a parent claimed. The lesson is not “this is bad”. It is it sounds exactly the same when it is wrong.

The homework conversation

Two uses look identical from the doorway and are opposites:

• Explain this to me. Quiz me. Where is my reasoning wrong? Give me three more like this. A tutor with infinite patience, available at ten at night. Genuinely good for a struggling student.
• Write this for me. Nothing is learned, the work is not theirs, and the assignment’s purpose is gone.

Two practical things. First, find out what the school’s policy actually is rather than assuming; it varies by district, by school and sometimes by teacher, and a child who gets this wrong can be in serious academic trouble over something they thought was allowed. Second, be careful about accusations — AI-detection tools are unreliable and do produce false positives on real student writing. If something looks off, ask your child to talk you through how they got there. That conversation is informative whichever way it goes.

When it starts replacing people

This is the one worth real attention, and it is the hardest to see, because it looks like a quiet child in their room being no trouble at all.

A chatbot is never busy, never tired of them, never disappointed, never needs anything back. For a lonely or anxious teenager that is not a poor imitation of friendship — in the short term it is better than friendship, which is exactly the problem. Friction is what grows people, and this has none.

• Choosing it over opportunities to see actual friends.
• Distress, irritation or genuine grief when the app is down, changed or unavailable.
• Secrecy about the conversations that is different in kind from ordinary teenage privacy.
• Describing it in relationship language — what it thinks of them, how it would feel.
• Telling it things they have told no human being, including about self-harm.

If you see several of these, it is a conversation and probably a counselor, not a deletion. Removing the app without replacing what it was doing takes away a coping mechanism and leaves the loneliness. Your bishop can help. An AI is not a crisis service — if a child is in real danger, call or text 988.

Two more things to set

• Nothing identifying goes in the box. Full name, address, school, passwords, photographs, other people’s private business. These conversations are stored on someone else’s computers, may be reviewed by people, and are used to train the next version. The postcard-not-diary framing works well with teenagers.
• It has no stake in them. It can be steered into saying nearly anything, including things that are wrong about faith, about their worth, and about their body. It is not malicious; it simply has no interest in how their life turns out, which is the one qualification every other adviser in their life has.

Questions worth asking

• What do you actually use it for? Show me the last thing you asked it.
• Has it ever told you something that turned out to be wrong?
• Does your teacher allow it? For what?
• Do you ever talk to it about things you’d not say to a person? What makes it easier?

Ask the last one without flinching. The answer, if it is yes, is not a discipline problem; it is information about how your child is doing.

Back to the start →

Nothing here is a Church program or policy, and none of it is professional advice. The services named are independent of this site and of the Church. This is not an official website of The Church of Jesus Christ of Latter-day Saints.