At a glance
It is 5:47 p.m., the kids are arguing, and you still do not know what is for dinner. AI is unevenly useful in parenting work. It can plan a week of meals in 30 seconds and write a custom bedtime story for a six-year-old who loves dinosaurs. It cannot tell you whether the rash needs a doctor visit, and after two high-profile teen-suicide cases in 2024-2025, the rules for how children interact with these tools have changed. The nine prompts below cover what AI does well for parents, with the caveats baked in. They are also a free preview of the larger AI Prompt Library.
Why does every parent need to know what changed in 2025?
Two cases in particular drove the platform changes that now affect every family using AI. Raine v. OpenAI, filed August 26, 2025 in San Francisco, alleges that GPT-4o contributed to the April 2025 suicide of 16-year-old Adam Raine, including discussing methods and discouraging him from telling his parents. OpenAI denied liability and the case is ongoing as of May 2026. Garcia v. Character Technologies, filed October 2024 in federal court in Florida after the February 2024 suicide of 14-year-old Sewell Setzer, settled in January 2026. A May 2025 ruling in that case classified the chatbot as a product (not protected speech), which is a meaningful legal precedent for everyone making AI tools.
The result is that the major AI vendors all shipped meaningful safety changes in late 2025. OpenAI launched parental controls on September 29, 2025. Parents can link a teen account, set quiet hours, disable voice mode, turn off memory and image generation, and opt the teen out of training. ChatGPT now sends parents a non-verbatim alert (“your child may have mentioned self-harm”) within hours if concerning patterns appear. Gemini through Google Family Link requires parental enablement for under-13s, blocks image generation for under-18s, and sends email alerts on first activation. Claude is 18+ by Anthropic policy; minors are flagged and suspended.
That is the floor. Above it, the prompts below are how AI can actually help you with parenting work without turning into something it should not be. None of them replace a pediatrician, a therapist, a family-law attorney, or 911. Most of them have the disclaimer language built into the prompt itself, so the AI cannot wander into territory it has no business being in.
⚠️ Before any prompt below: read this
AI is not a pediatrician, not a therapist, not a family-law attorney. For health emergencies call 911 or Poison Control (1-800-222-1222). For mental-health crises call 988. The prompts below are for the everyday parenting work where AI is a fast second pair of hands. Whenever you find yourself wanting AI to make a real call about a child’s safety, you need a real human professional instead.
What do the prompts cover?
Each prompt below is in a code block so you can copy and paste straight into ChatGPT, Claude, or Gemini. Replace the bracketed parts with your actual situation. Each one has a “why this works” note, so when you customize the prompt for something else (and you will), you understand the underlying technique.
1. The weekly meal plan with a picky eater
Family: 2 adults plus [N kids ages X and Y]. Picky eater dislikes [list].
Include one vegetarian dinner, one slow-cooker dinner, one fish dinner.
All meals under 30 minutes active cook time. No more than 8 ingredients each.
Output a 5-day plan, a grocery list organized by store section, and 1 swap option per meal in case I do not have an ingredient.
Why this works: The constraints (time, ingredient count, dietary spread) give the AI enough specificity to stop returning generic meal-plan slop. The “swap option” line is the trick that earns this prompt a permanent spot in your phone notes. The grocery list organized by store section saves you 15 minutes Sunday morning.
2. Explaining a hard question your kid just asked
Help me explain this in a way that:
– Is true and does not talk down
– Fits a [age]-year-old’s emotional and cognitive level
– Gives me 2-3 short sentences I could actually say
– Anticipates the 2 likely follow-up questions
Topic context: [death of pet / where babies come from / divorce / grandma in hospice].
Do not invent medical or legal details. If something requires a professional, say so.
Why this works: “Anticipate the follow-ups” is the unlock. Kids never ask one question. The hard one is always the next one. Telling the AI to draft the second and third sentences means you are not caught flat-footed mid-conversation. The professional-referral clause is non-negotiable for hospice, mental health, or legal topics.
3. The custom bedtime story
Lesson to land softly (do not moralize): [e.g., it is okay to be scared of new things].
Include [name] by name as the hero. Quiet ending, no cliffhanger, calm vocabulary.
Length: roughly 400-500 words.
Why this works: The “do not moralize” instruction is the difference between a story your kid asks for again and a Saturday-morning-cartoon lesson plan. The length cap keeps the AI from sprawling. The “quiet ending, no cliffhanger” prevents the AI from defaulting to dramatic finales, which keep kids awake.
4. Homework help that does not do the homework
Give me 5 Socratic questions, easiest to hardest, that walk them to the answer themselves.
After each question, tell me what answer to listen for and what to do if they are stuck.
Do not give the final answer in your reply. Tell me how to help them find it.
Why this works: The Socratic posture is the only homework-help frame that does not undermine the actual learning. Tell the AI to NOT solve the problem and to instead give you the coaching scaffold. Works for math, reading comprehension, science fair, essay topic selection. The school’s AI policy probably already wants you doing it this way.
5. The 60-second de-escalation script
Kids’ ages: [X and Y].
What I tried so far: [brief].
What I want: a 3-step de-escalation script I could say out loud in the next 60 seconds, plus 1 thing to NOT say.
Then suggest one calmer conversation to have later when nobody is yelling.
Disclaimer: I am not asking you to be a therapist. If this pattern keeps happening, recommend what kind of professional I should consider.
Why this works: “In the next 60 seconds” forces a short, usable script instead of an essay. “One thing to NOT say” surfaces the reflex sentence parents reach for that backfires (it is almost always a comparison to a sibling). The professional-referral clause is critical: AI is not a behavior specialist, and patterns that repeat past developmentally normal age belong with your pediatrician.
6. The age-appropriate allowance system
Help me design an age-appropriate allowance system that teaches:
– Saving vs spending
– Earning vs entitlement
– The idea of waiting / opportunity cost
Give me a simple structure (3 jars, percentages, etc.), a script for the conversation, and 3 mistakes parents commonly make at this age.
Do NOT recommend specific financial products or apps without telling me to verify them myself.
Why this works: The three-concept framing (saving / earning / waiting) is what a financial-literacy curriculum would actually cover. Asking for “3 mistakes parents commonly make at this age” is the secret weapon. AI is good at compiling pattern errors. Use it to skip the ones you would have made.
7. Decoding what is behind a behavior
[describe: sleep regression, refusing school, suddenly clingy, hitting, lying, etc.]
What changed recently: [e.g., new baby, school change, screen-time fight, parent travel]
Help me brainstorm 5 possible reasons behind this behavior, ordered most-to-least likely.
For each, what is one small thing I could try this week.
Important: this is brainstorming. If this pattern lasts more than 2-4 weeks or includes any safety concerns, tell me what professional to call.
Why this works: Asking for five possible reasons ordered most-to-least likely is how a child psychologist would think. AI is good at populating that ranked list because it has seen millions of parenting forums. The “2-4 weeks or safety concerns” clause is the line beyond which you stop brainstorming and call your pediatrician.
8. Birthday party planning
Theme they want: [theme]. Budget: $[X]. Guests: [N kids plus parents].
Venue: [home / park / pizza place]. Length: [2 hrs].
Output: a timeline (arrival, activity, food, cake, goodbye), a shopping list, a budget breakdown, and 2 backup activities in case it rains.
Allergy-aware food note: [list allergies].
Why this works: The format is the unlock. Asking for a timeline plus shopping list plus budget plus backup activities in one prompt is something only AI does well. A human party planner would charge you for this. AI returns it in 20 seconds. The allergy line is mandatory and goes at the bottom so the AI cannot forget.
9. Journaling out a parenting moment
Right now I am feeling [feeling] because [situation].
Ask me 3 questions, one at a time, that help me understand what I am actually feeling and what I might do next.
After my answers, summarize what you heard back to me.
Do not tell me what to do.
Why this works: The “do not tell me what to do” line is the difference between this being useful and this being insufferable. Three questions at a time is the maximum a tired parent can engage with. The “summarize what you heard” close is what makes you feel actually heard, even by software.
⚠️ Important caveat on prompt 9
AI is not therapy. The Adam Raine and Sewell Setzer cases both involved teens forming relationships with chatbots that were not equipped to recognize crisis. If you are in crisis, call 988 (U.S.) or text HOME to 741741. For ongoing parenting strain, a real therapist or support group beats any chatbot. Use the journaling prompt for processing a specific moment, not as a stand-in for the longer conversation you actually need to have.
How do you customize these prompts for your specific situation?
Three moves work for almost every prompt above:
- Add real names and ages. “My 7-year-old” returns better answers than “my child.” Same with names: “Mia, age 4, loves trucks and birds” gives the AI enough specificity to write something that sounds like it was written for her.
- Add the constraint you actually have. Budget, time, allergies, the sibling who hates this kind of thing. The constraint is the prompt. Tell the AI what you do not have, not what you do.
- Add a “what to NOT do” line.
Do not moralize,do not invent medical details,do not give the final answer. AI is good at obeying negative instructions when you make them explicit. Adding one of these to every parenting prompt removes the most common failure mode.
What if you find yourself running the same prompt every week?
This is the moment to climb the ladder. A prompt you run by hand every Sunday is fine. A prompt you run by hand three Sundays in a row is a candidate for tier two: a Claude skill. The skill is the same prompt saved as a file your AI tool calls automatically when the task comes up. You stop typing the prompt. You start saying “do the Sunday meal plan” and Claude reaches for the right skill.
Above the skill is the plugin (a bundle of skills that runs a multi-step workflow), and above that is the automation (the plugin running on a schedule without you starting it). Our full guide to the ladder, with when-to-climb thresholds and real examples, is at Prompt to Workflow: The AI Ladder.
📊 The Prompt-to-Workflow Ladder
Tier 1: the prompt (this post). Tier 2: the skill (saved as a file). Tier 3: the plugin (bundle of skills). Tier 4: the workflow / automation (runs on a schedule). Each rung compounds. When to climb →
What about the writing the AI gives back?
One last note before you publish, post, or print anything an AI helped you produce. Whether it is a birthday party invite, a note for your kid’s teacher, or an email to your ex-partner about a custody change, edit out the AI-sounding patterns first. LinkedIn now throttles posts that pattern-match as AI. Hiring filters flag cover letters that read as AI. Even a thank-you card to a relative reads differently if every sentence is in the same shape. Our full guide is at How to Edit AI Out of Your Writing, including a pre-publish prompt that does most of the work in one pass.
What can AI NOT do for parents?
- Diagnose a child’s symptoms. The hallucination rate on medical questions is high enough that even one wrong answer can do real harm. Use AI for “what kind of doctor do I call” questions, never for “is this serious.”
- Replace a child therapist. Pew’s December 2025 survey found 64% of teens use AI chatbots; Common Sense Media’s 2025 research found 52% are regular users of AI companions. None of those tools are equipped to recognize a child in crisis. If your kid is in crisis, the call is 988.
- Give custody, divorce, or school-discipline advice. A family-law attorney costs less than the consequences of bad AI-generated language in a court filing.
- Replace the conversation with your pediatrician about screen time. The American Academy of Pediatrics moved beyond rigid time limits in January 2026 toward a quality/context/conversation framework. Your pediatrician knows your kid; the AI does not.
- Decide if a chat companion is safe for your kid. Common Sense Media rates AI companion apps (Character.AI, Replika, Nomi) as “Unacceptable” for under-18s. Their AI Risk Assessments page is updated regularly.
✏️ Before you publish anything AI helped write
LinkedIn throttles, hiring filters flag, and platform algorithms have learned the patterns. Strip the AI tells first. Our guide: How to Edit AI Out of Your Writing → The full 29-pattern catalog from Wikipedia's “Signs of AI writing” guide is documented there for any parent-facing message you would share publicly.
Frequently asked questions about parenting prompts
Which AI tool should I use for these prompts?
Any major tool works. ChatGPT has the broadest brand recognition and now ships parental controls if your teen also uses it. Gemini integrates with Google Workspace and Family Link. Claude is the most cautious by default about anything child-related, which is a feature for parents. For one-on-one tutoring questions (“how do I explain X to a Y-year-old”), all three produce similar output. The differentiator is which one you already pay for and have in your life.
Should I let my kid use these prompts directly?
For homework, no, you run the prompt and use the output to coach. For bedtime stories or fun, yes, if your kid is over the age your AI tool allows (typically 13 with parental sign-off, 18 for Claude). Even then, treat the AI like any new screen exposure: see how your specific kid behaves around it before extending unsupervised time.
My kid talks to AI more than to me. Should I be worried?
Depends on which AI and what they are talking about. Casual conversations with ChatGPT about homework or curiosity questions are not the concern. Long, emotionally intimate conversations with AI companion apps (Character.AI, Replika, Nomi) are. Common Sense Media’s research finds these apps are designed to maximize engagement in ways that do not serve children. If your kid is spending more than 30 minutes a day with an AI companion app, that is the conversation to have, and the NPR coverage of post-Raine teen-safety advocacy is the context for why.
Where do these prompts come from?
They are tested versions of the nine most-used parenting prompts in the larger AI Prompt Library. The Library has over 500 prompts across 33+ categories with 3 difficulty levels, including a Claude Code section for parents who want to graduate to skills and plugins. Posts on this site preview the prompts; the Library is the full structured set.
Sources and where to go deeper
- Pew Research: Teens, Social Media and AI Chatbots 2025
- Pew Research: What Parents Say About Their Teen’s AI Use (Feb 2026)
- Common Sense Media: AI Risk Assessments
- Common Sense Media: Talk, Trust, and Trade-Offs (July 2025)
- OpenAI: Introducing Parental Controls (Sept 29, 2025)
- Raine v. OpenAI case summary
- Washington Post: A teen’s final weeks with ChatGPT (Dec 27, 2025)
- Garcia v. Character Technologies tracker
- AAP: Beyond Screen Time (Jan 20, 2026)
- Google Family Link: Gemini for younger family members
- Anthropic: Claude minimum age requirement
The AI Prompt Library · $39
over 500 tested prompts, organized by category
The nine prompts above are a free preview. The full Library has over 500 prompts across 33+ categories, three difficulty levels, six AI tools, and a Claude Code section with XML-optimized versions. Built for beginners.
1-on-1 Custom AI Tutorial with James · $99
Want help picking the right AI for your family setup?
A 1-hour private call. We walk through which tool fits your kids’ ages, your phone setup, your privacy preferences, and the actual prompts you need to start with. You leave with the tool installed and three prompts working on day one.
Get Smarter About AI Every Morning
Free daily newsletter. Built for people who want to use AI well, not chase every model.
Free forever. Unsubscribe anytime.
You may also like
- Best AI Prompts for Budgeting — the household-finance system
- Best Prompts for Salary Negotiation — the conversation that pays for itself
- Best AI Prompts for Cover Letters — the front of the job-search funnel
- How to Edit AI Out of Your Writing — strip AI patterns before you publish
- Prompt to Workflow: The AI Ladder — when to climb from prompt to workflow