Don’t hire AI as your bedtime lawyer

Kids are growing up fast. They have already found one of the most advanced uses of artificial intelligence: a second opinion when Mum says bedtime.

You say “lights out.” They open a chat. Two minutes later they return with a very reasonable paragraph about sleep cycles, weekend exceptions, and how “some families” do 9:30. The tone is polite. The goal is a stay of execution.

It is funny until you notice the pattern. The machine is being hired as a lawyer.

The new appeal

Bedtime arguments are old. The new bit is the exhibit. A chatbot will happily produce a plan that sounds official: wind-down steps, a later clock “for older children,” a note that one more chapter supports literacy. Some of that may even be useful on a Sunday afternoon, when you are building a routine on purpose.

At 8:47 on a school night, it is not research. It is a negotiation tactic with better grammar.

You do not have to mock them. Curiosity is a good skill. The mistake is letting the chat sit in the adult chair. The model does not know who has swimming at 6:30, who has a test, who has already had two delays, or who is the person that will be tired in the morning. You do.

A house rule that fits on a sticky note

Write it once. Stick it near the charger.

You may ask AI for ideas. You may not hire it as your bedtime lawyer.

That split keeps the useful part. On a calm afternoon, a child can ask: “Give our family three simple wind-down ideas that take ten minutes.” You look at the list together. You pick one. You try it for a week. That is AI as a suggestion box.

The lawyer version is different. It arrives after the decision. It treats the parent as the opposing party. It asks the chat to find a loophole.

When that happens, you can smile and still hold the line:

• “Thanks for the research. The answer is still bedtime.”
• “Ideas on Saturday. Decisions tonight are ours.”
• “AI does not live here. We do.”

You are not anti-science. You are anti-appeal-court-in-pyjamas.

What this teaches besides sleep

If a chatbot can overturn bedtime, it will be asked to overturn screen limits, vegetable policy, and “you still have to apologise.” The lesson underneath is rank: who decides, and when a tool is allowed to speak.

Let them keep the skill of asking for ideas. Take back the job of deciding. Children need both — a way to gather options, and a grown-up who can say “that’s enough for tonight” without a citation.

If the routine itself is a mess, fix the routine in daylight. Pick a time. Pick a wind-down. Write it where everyone can see it. Then the chat has nothing to win at 8:47.

If they keep appealing, do not debate the science of melatonin in the hallway. Repeat the rule once. Offer a choice that is still a bedtime: which book, which cup, which stuffed animal sits on the pillow. Choice inside the boundary. Not a new hearing.

Some nights you will give an extra ten minutes because the day was hard. That is parenting, not a precedent the chatbot can cite tomorrow. Say so: “This is a tonight exception. It is not a new time.”

AI can help a family design a bedtime. It does not outrank the people responsible for the child. Kiss, water, light off. The lawyer can wait until morning.