Download the Word review copy (.docx)
Homework Help or Homework Replacement?
Chapter 4 sample from the planned guide Your Kid Is Using AI. Now What? You don’t need to understand every AI tool to have a useful conversation.
An adult-facing resource for parents and caregivers supporting children ages 10 to 15. No child account or chatbot use is needed to try the activity.
Preview 0.11 · 20 October 2026. For adult review. Professional review is pending, and some claims still need verification. Product rules can change. General educational information, not medical, legal or emergency advice.
The situation
Maya is 12. She has a science paragraph due on how volcanoes form. Her dad, Chris, sees a chatbot open on her laptop.
Chris: “Are you using AI to do your homework?”
Maya: “I only asked it to explain volcanoes. Then I asked for an example paragraph. I changed the words.”
Chris: “That sounds like cheating.”
Maya: “You use AI at work all the time. That’s not cheating.”
Now they are both upset. Neither is sure where the line is, and nobody has checked what her teacher said. You do not need to be a tech expert to handle this. You need a few good questions and a plan.
Before they go further, Chris also needs to check whether Maya is eligible to use that specific service. They can work through the rest of this chapter without opening a chatbot.
Check age rules first
Look up the rules for the exact tool your child wants to use, and check the date on what you read. Do not help a child get around an age rule, and do not share one account to skip it. See sources
Product policy or feature Google allows some children under 13, or the applicable age in their country, to use Gemini Apps through a supervised account when a parent enables access in Family Link. Availability varies by region and account. Check Google’s current eligibility information before use.
Suggested family practice Being eligible to use a tool does not mean it is suitable for every child or every task.
Why this is hard
AI can help with homework in very different ways. The tool is the same, but the job is not.
- Helping you learn. Explaining an idea a new way. Quizzing you.
- Helping you start. Brainstorming. Suggesting an outline.
- Doing the part you should practice. Writing the paragraph. Giving the answer to copy.
Three checks to use every time
- Learning check. What is this assignment supposed to teach? Is the tool helping me learn that, or doing it for me?
- Teacher check. What does my teacher allow? Brainstorming may be okay. Turning in a generated answer may not be.
- Honesty check. Can I say clearly what I did and what the tool did?
Product policy or feature AI can produce an answer that sounds convincing but is factually wrong. Some providers call these mistakes “hallucinations.” Confidence is not proof that an answer was checked. Suggested family practice When a claim matters, open a reliable source and see whether it supports or contradicts the claim. See card 4
Product policy or feature Some chatbots can search the web and link to sources. That does not guarantee the answer is right. Open the source and check whether it actually supports the claim. See sources
Product policy or feature What is saved, reviewed, or used to improve a service depends on the product, account type, settings, and whether you submit feedback. Check the product’s current policy. Suggested family practice A family boundary to consider: do not upload passwords, identifying school information, private conversations, or another person’s photos or details. Use made-up examples when possible. See card 6
What to say
Start with curiosity, not a verdict. Here is how Chris could replay the moment.
Chris: “I’m not assuming you did something wrong. What did you ask it to do?”
Maya: “Explain volcanoes. Then write a sample paragraph.”
Chris: “Thanks for telling me. An explanation might help. Let’s see what you understood and what the assignment asks you to practice. The sample paragraph is the part I’m not sure about.”
Maya: “How to explain something in my own words.”
Chris: “Then let’s check what your teacher allows. And you asked a fair question about my work. Let’s talk about my rules, too.”
Chris thanked her, asked questions, named the learning goal, and promised to check with the teacher. He did not need to know everything. Try card 3
What to check with the teacher
Research finding School AI rules can be unclear. In a 2026 U.S. survey of 1,017 teens ages 13 to 17, 35% said they did not understand their school’s AI rules or were unsure what they were. See sources
Suggested family practice Check the directions for the specific assignment. If the rules are unclear, ask the teacher what is allowed and how AI use should be acknowledged.
Look for written rules first, like the syllabus or class page. If they are not clear, ask:
- Can students use AI for this assignment? For which parts?
- Is brainstorming or getting an explanation okay?
- Do I need to say when I used it? How?
- What would count as going too far?
Hi, my child has been using an AI tool for schoolwork and we want to follow your rules. Is it okay for explaining ideas or brainstorming? Is it okay for drafting? How should my child tell you when it was used? Thank you!
Teacher permission is for the task, not for everything. It also does not change a tool’s own age rules or privacy settings. See card 2
Try it together: Sort and Explain
About 20 minutes. You need a recent assignment’s directions (not private chats), paper, and a pen. Your child does not need to have used a chatbot. This works either way.
- Name the goal. Read the directions together in a suitable way, including established accommodations. Finish this sentence: “This assignment helps me learn how to ___.”
- Make three columns: “Helps me learn,” “Ask the teacher,” and “I need to do this myself.”
- Sort. Write ways AI could be used, like “Explain the topic,” “Quiz me,” “Write my first sentence,” “Write the whole thing.” Unsure? Put it in the middle and ask the teacher.
- Show understanding. Pick one thing your child learned for this assignment, from class notes, a book, a chatbot or another source. They can speak, write, draw or use another suitable method, including established accommodations and learning supports. Fluent speech or unaided performance is not required. Difficulty explaining is not proof of AI use. Explore what they understand and what support or practice would help.
- Write one family rule. Example: “We use AI to understand, not to replace the part we’re supposed to practice. When we’re not sure, we ask the teacher.”
- Include yourself. Your child asks: “When do you use it for work, and what do you check?” Answer honestly.
Your child does not have to show private chats. A short description is enough. Open the checklist
Notes by age
Ages 10 to 12. Keep rules simple and stay close. Check each tool’s age rules before any use. Sit together at first and invite them to show understanding by speaking, writing, drawing or another suitable method, including established accommodations.
Ages 13 to 15. Bring them into writing the rule. Talk about honest credit: what I did, what the tool did. Check age and parent-permission rules for each product. Use a suitable way to show understanding, including established accommodations; verbal fluency or unaided performance is not a test for AI use.
If something goes wrong
Review agreed, not completed: A mental health professional has agreed to review the Chapter 4 sample and the first batch of conversation cards. That agreement does not cover the full guide, checklist or website. No completed review, approval or endorsement is claimed.
If your child broke a class rule with AI, stay calm and thank them for telling you. Then contact the teacher together and ask how to fix it.
If your child saw something upsetting, they do not have to show it to you to get help. Stop, listen, and decide the next step together.
Six conversation cards
Change the words so they sound like you.
Card 1: How is this different from a calculator?
That’s a useful comparison. The question is what the assignment is meant to help you learn. Sometimes a tool supports that learning. Sometimes it does the part you’re supposed to practice.
What is this assignment supposed to help you learn?
Read the assignment directions together and check what your teacher expects.
Note for parents. A calculator can be okay on one task and not allowed on another. The same goes for AI. This comparison helps start the talk, but it does not settle every case.
Card 2: What if my teacher says it’s okay?
Then let’s understand exactly what’s allowed. Permission for brainstorming might not mean permission to turn in a generated answer.
What exactly did your teacher say you can do?
Find the rule in writing, or ask the teacher to explain it.
Note for parents. A teacher’s okay covers schoolwork rules. It does not change a tool’s age rules or privacy settings. Check those separately.
Card 3: Why can you use AI for work, but I can’t use it for homework?
That’s a fair question. I don’t want to make a rule I can’t explain. Let’s figure out what you’re supposed to learn, what your teacher allows, and which parts you need to do yourself. We can also talk about the rules I should follow.
Show me what you wanted help with.
Sort the task into “helps me learn,” “ask the teacher,” and “I need to do this myself.”
Note for parents. Be ready to answer honestly about your own rules at work, including what you check and what you never share.
Card 4: Why would it make something up?
It can write an answer that sounds right even when the facts aren’t there. A confident answer doesn’t mean it checked.
Which part of this answer would you want to double-check?
Pick one claim from the answer. Look in a reliable source to see whether it supports the claim or contradicts it. Don’t just look for something that agrees.
Note for parents. Some tools can look things up, but the answer can still be wrong. Checking a source is a skill worth practicing.
Card 5: Why should I believe you instead?
You don’t have to assume I’m right either. Let’s look for good evidence and check both answers.
What would help us decide?
Choose one question you disagree on. Each of you finds a source. Compare what they say, and check whether each source supports or contradicts each answer.
Note for parents. If safety is at stake, such as a stranger asking for personal information, act first and look into it together afterward.
Card 6: Can someone read what I type?
We should not assume a chatbot conversation is private. What is saved, reviewed, or used to improve the service depends on the product, account type, settings, and whether you submit feedback. Let’s check before sharing anything personal.
What kinds of things do you usually type into it?
Check the product’s current policy. Open the privacy settings together and find what is saved. A family boundary to consider: Do not upload passwords, identifying school information, private conversations, or another person’s photos or details. Use made-up examples when possible.
Note for parents. Turning off “use my chats for training” does not always mean chats are not saved. Some may still be kept or accessed for safety or legal reasons. Read the service’s own description, and check its age rules first. Product-specific check dates and limits are in the sources. See sources
Homework checklist
This is a conversation guide, not a test for catching your child using AI. Use it together. Your child can read the “I” statements, and you can help check school rules, tool eligibility, and privacy.
Ticks and your family rule are saved in this browser when available; this checklist does not send them to the Site server. Browser entries do not have automatic 30-day expiry. The clear buttons reset the ticks or empty the rule in this browser only; they do not delete submitted feedback. Avoid entering names, school details, or private conversations. Other people using this browser may be able to see saved entries.
Saving is not available in this browser, so your ticks and rule will not be kept.
Before they start
While they work
When they finish
Questions for the teacher
Our family rule
To report a mistake, contact on LinkedIn.
Sources and checks
Source details
Each note shows what the source supports and where it stops. Open the source before relying on it.
Teens in the AI Era: Schoolwork and Skills That Matter Research finding
Publisher: Common Sense Media (survey by NORC at the University of Chicago). Published: 2026. Checked: 18 October 2026 (inherited).
Supports: 35% of teens did not understand their school’s AI rules or were unsure of them. 23% said AI was banned at their school.
Where: Key finding 7, printed page 9
Applies to: US teens ages 13 to 17 (1,017 surveyed April 30 to May 14, 2026). It does not cover ages 10 to 12.
Limit: Self-reported survey. Margin of error about 4.3 points overall, and higher for subgroups.
Guide your child’s Gemini Apps experience Product policy or feature
Publisher: Google For Families Help. Published: Date not shown on the page. Checked: 19 October 2026 (independent page read).
Supports: A child under 13 (or the applicable age in their country) can use Gemini Apps with a supervised account only if a parent enables access in Family Link. The Keep Activity setting is not available to them. Gemini can hallucinate and present inaccurate information as factual.
Where: Supervised account eligibility, Privacy, availability and parent enablement sections
Applies to: Gemini Apps on supervised accounts. Not available for supervised accounts in the EEA, Switzerland, or the UK, and rolling out gradually.
Limit: It does not say nothing is stored or reviewed, and it does not cover personal teen accounts or school-managed accounts.
Why language models hallucinate Vendor research
Publisher: OpenAI. Published: September 2025. Checked: 18 October 2026 (inherited).
Supports: OpenAI describes hallucinations as plausible but false statements and says they remain a challenge for all language models.
Where: Opening section of the page
Applies to: Language models in general, as described by one vendor.
Limit: It explains why mistakes happen and says newer models have lower rates. It does not rate any tool for homework.
Searching the web with ChatGPT Vendor documentation
Publisher: OpenAI Help Center. Published: Date not shown on the page. Checked: 18 October 2026 (inherited).
Supports: ChatGPT can search the web and link to sources. Results and citations can be incomplete, outdated, or incorrect, so open the source and check it.
Where: Overview, and “Review sources and results”
Applies to: ChatGPT web search only. Other chatbots may differ.
Limit: It does not mean every response uses search.
Manage & delete your activity in Gemini Apps Vendor documentation
Publisher: Google Gemini Apps Help. Published: Date not shown on the page. Checked: 18 October 2026 (inherited).
Supports: With Keep Activity off, conversations can still be saved with the account for up to 72 hours to provide the service and process feedback.
Where: “Turn off Keep Activity”
Applies to: Personal Google accounts. Work or school account settings are controlled by the administrator.
Limit: It does not describe supervised-child or school-managed accounts.
Gemini Apps Privacy Hub Vendor documentation
Publisher: Google Gemini Apps Help. Published: Privacy questions updated October 13, 2026 on returned page. Checked: 19 October 2026 (independent page read).
Supports: Chats reviewed by human reviewers, and related data, can be kept for up to three years and are not deleted when you delete your activity. With Keep Activity off, Google still uses chats to respond and to protect its users.
Where: Retention, human review and Keep Activity off sections
Applies to: Gemini Apps when signed in. Work or school accounts may have different terms.
Limit: It describes specific circumstances, not every chat.
Artificial Intelligence and the Future of Teaching and Learning: Insights and Recommendations Policy guidance
Publisher: U.S. Department of Education, Office of Educational Technology. Published: May 2023. Checked: 18 October 2026 (inherited).
Supports: It recommends keeping humans in the loop, and says teachers should remain at the helm of major instructional decisions.
Where: “Key Recommendation: Human in the Loop AI,” printed page 17
Applies to: US K-12 policy context.
Limit: General guidance, and the report says it does not have the force of law. It is not your child’s school policy. The Department’s core-messages handout could not be opened on October 18, 2026.
Is ChatGPT safe for all ages? Vendor documentation
Publisher: OpenAI Help Center. Published: Date not shown. Checked: 19 October 2026 (independent page read).
Supports: ChatGPT is not meant for children under 13, and ages 13 to 18 need parental consent.
Where: Full article opened; minimum age and parental consent bullets
Applies to: ChatGPT consumer use.
Limit: Provider age rules, not a finding that a tool is suitable or safe for every child.
Consumer Terms of Service Vendor terms
Publisher: Anthropic. Published: Effective October 8, 2025 on returned page. Checked: 19 October 2026 (independent page read).
Supports: The returned consumer terms require users to be at least 18.
Where: Account creation and access, minimum age
Applies to: Returned terms for consumers resident in the EEA or Switzerland; check the version applicable where you live.
Limit: Regional terms page; no claim that every regional version was checked.
Data controls in ChatGPT Vendor documentation
Publisher: OpenAI Help Center. Published: Date not shown. Checked: 19 October 2026 (independent page read).
Supports: Turning off model training does not delete saved chats. New chats can remain in history.
Where: Training toggle and saved chats sections
Applies to: ChatGPT consumer accounts.
Limit: Controls depend on account, plan and workspace; not an audit of all retention paths.
Google Gemini: What these retention periods mean
Product policy or feature Google’s personal-account documentation says that even when Keep Activity is off, conversations can be saved with the account for up to 72 hours to provide the service and process feedback.
Separately, Google says chats reviewed by human reviewers, along with related data, can be retained for up to three years and are not removed simply by deleting activity.
These describe different circumstances. They do not mean every chat is kept for three years, or that turning activity off immediately deletes everything. Do not assume personal-account settings describe a school-managed or supervised-child account.
Google also says the Keep Activity setting is not available to users under 13, or the applicable age in their country. That fact alone does not establish that nothing is stored or that nobody can review a conversation. Check the privacy information for the specific account and service. The Privacy Hub retention passages and supervised-child eligibility page were independently checked on 19 October 2026; the separate activity-management page remains an inherited 18 October check.
More from the school survey
Research finding Among surveyed teens who used AI for schoolwork and had an AI tool blocked on a school network or school-issued device, 59% said they used a personal device to access AI instead. This applies to that subgroup, not all students. A block is not the same as breaking an assignment rule.
Policy guidance A 2023 U.S. Department of Education report recommended keeping humans involved in educational AI decisions, including teachers retaining responsibility for major instructional choices. This is general policy guidance, not your child’s assignment policy.
Still awaiting verification
- Whether chatbots other than ChatGPT can search the web. The note above covers ChatGPT only.
- Provider pages outside the five independently opened on 19 October 2026 have not been rechecked for changes.
- The Department of Education core-messages handout, which could not be opened.
- The five checks below cover the cited passages, not a legal, privacy or product safety audit.
Left open on purpose
School and district AI rules vary, so this toolkit does not set one. Check the assignment directions and ask your child’s teacher or school what is allowed.
Passages that need qualified review
- The help-seeking commitment and “If something goes wrong” (safeguarding and mental health).
- Privacy wording on card 6 and in these sources (privacy).
- Age and account statements (legal).
This is a draft. A mental health professional has agreed to review the Chapter 4 sample and the first batch of conversation cards. That agreement does not cover the full guide, checklist or website. No completed review, approval or endorsement is claimed. Five official provider pages were independently opened on 19 October 2026 for the stated claims; other checks remain inherited from version 0.5.
Preview provenance and review scope
Preview 0.11, 20 October 2026. Adapted from the unchanged Claude version 0.5, 18 October 2026. Draft for parent testing. A mental health professional has agreed to review the Chapter 4 sample and the first batch of conversation cards. That agreement does not cover the full guide, checklist or website. No completed review, approval or endorsement is claimed. Teacher, privacy, safeguarding and legal review have not been completed. Source notes near the end distinguish five provider pages independently checked on 19 October 2026 from inherited 18 October checks. A few claims are still awaiting verification, as listed in the sources section. Provider rules change, so check each product’s current rules. General educational material, not medical, legal, or emergency advice.