Guidance In-House Third-party AI Mobile App
Guidance

Using AI at YCYW

AI can support learning only when human thinking and values stay at the centre.

Explore the guidance below to learn how you can use AI in an intentional and ethical way to enhance your learning.

1 / 10 2 / 10 3 / 10 4 / 10 5 / 10 6 / 10 7 / 10 8 / 10 9 / 10 10 / 10
The question that changes everything
Don't ask Can AI do this for me?
Ask What should I think through myself, and what should I ask AI to help with? And why?
You are in charge of the thinking. AI is a tool you tell what to do.
The Thinking Cycle

Four steps to work through on any task. Step 1 is where you decide which thinking you will do yourself.

1 I think first Work out what you already know, and which parts you need to think through yourself.
2 AI thinks Ask AI to help with the parts you chose to hand over. Check what it gives you.
3 We think together Compare what AI said with your teacher, your classmates and other sources. The final work is yours.
4 We think better Ask yourself what you understand now that you did not understand before.
Step 4 is not just the end. Keep asking what you understand better, at every step along the way.
1. I think first Executive control. What do I know already? Which parts need my reasoning? Which sub-tasks can I hand over, and why?
2. AI thinks Cognitive outsourcing. Clear prompts, bounded tasks. AI can be wrong, biased, and can make things up. Do not let it cross the line you drew.
3. We think together Distributed cognition. What was useful? What was wrong? How does it compare with your teacher, your partner, your experiment? The final product must be yours.
4. We think better Metacognitive reflection. What did I keep? What did I delegate? Did my understanding deepen, or did I let AI think for me?
How much AI is allowed on this task Every task tells you how much AI you may use. Your teacher sets this, or you agree it together.
Red No AI. Do this yourself.
Amber You may use AI for certain parts. Say what you used it for.
Green AI is part of the task. Show how you judged what it gave you.
Amber 1 · Ideas only AI for brainstorming or outlining. Final product without AI.
Amber 2 · Feedback Your draft first. AI for feedback only.
Amber 3 · Selective tool AI for specific components only. You decide what, and why.
The test

Ask yourself this when you have finished a piece of work.

Can you explain your work, and why you made those choices, right now, without notes or AI?
YES AI enriched your learning.
NO You let AI think for you. Go back to Step 1 and try again.
By age

How much students use AI on their own, at each stage. Your school sets the detail.

Ages 0–8 Play and stories No independent use of generative AI.
Ages 8–11 With an adult Approved tools only, with a teacher or parent present.
Ages 12–14 With guidance The four steps are taught directly.
Ages 15–18 With autonomy Judgment expected. Defend conclusions without AI.
Ages 8–11 Three questions to practise: What should I try myself first? What can AI help me with? Does this answer make sense? Observations, predictions and conclusions stay with the student.
Ages 12–14 Growing independence. AI-enhanced inquiry begins in science and other inquiry subjects.
Ages 15–18 Stays with the student: hypothesis, experimental design, data interpretation, argument, evaluation of evidence. Learn to detect bias, hallucination and overgeneralisation.
For you

The same four steps, written for the part you play. Choose yours.

Think first Write your own rough ideas before you open any tool. Decide what stays yours.
Check everything AI makes things up with confidence. Compare it against a second source.
Say what you used Tool, date, prompt, and what you changed. If AI helped at all, say so.
30 seconds after What did I learn? Did my thinking get stronger or weaker? What next time?
Six things only you can do AI can help with plenty. These six stay with you, and they are the parts that make the work yours.
Feel curious Wanting to know why is yours alone.
Decide what is right Only you can weigh what matters.
Build friendships Trust is made between people.
Take responsibility You stand behind your work.
Know what matters to you Your values are your own.
Create meaning You decide what the work is for.
A runner does not get fitter by riding in a car.
Set the level State Red, Amber or Green on every task. Say which amber level.
Ask for evidence Initial thinking, what was delegated and why, the product, a short reflection.
Talk about AI use in feedback Show where AI enriched learning. Discuss, anonymously, where it did not.
Keep a balanced diet Red builds independence. Amber builds judgment. Green pushes inquiry.
Across the teaching cycle
Before class Adapt resources and differentiate. Explore other ways to explain.
During class Live differentiation, formative feedback, simulations. The teacher leads.
After class Targeted feedback and pattern-spotting, with human oversight. Tell students when AI grading is used.
The greatest risk is not dishonesty. It is students who stop thinking because AI is available.
Ask one question “Can you explain this to me without your notes?” That is the whole test.
Under 11, sit with them Approved tools only, with an adult present. Under 8, no independent use.
Nothing private, ever No names, addresses, photos or private feelings. Treat every chat as public.
Watch for reliance If a chatbot is becoming company, talk to us. Human relationships come first.
If something goes wrong Tell a trusted adult at school straight away. This includes disturbing content, and AI used to bully or impersonate someone. Your school's Designated Safeguarding Lead is there to help.
Staying safe

Three things everyone should know before using any AI tool, at school or at home.

Three things to remember
AI is not private What you type may be stored and read. Treat every conversation as public.
AI is not your friend It does not know you or care. If you are turning to it for support, talk to a trusted adult.
Some tools want your time Set limits. Close the tool when you have what you need.
Citing AI

Follow your teacher’s required citation style and AI-use policy. APA lists the AI developer as the author; MLA begins with a title or description of the generated content and lists the developer as publisher. Explain how you used AI as well as providing any required citations.

Your AI-use disclosure, include these four things:

1 Name the tool and model Record the tool and the actual model name if shown. Do not guess a model or version.
Example Claude Express, on the YCYW AI platform; model: {actual model name, if shown}
2 State the date AI answers change over time, so the date matters.
Example 4 September 2026
3 Include the prompt Write out what you actually asked it.
Example “Give me three counter-arguments to my conclusion that the experiment was reliable.”
4 Say what you used it for Name the part AI helped with, and the thinking you did yourself.
Example I asked for counter-arguments after writing my own conclusion. I chose two, checked them against our class data, and rewrote my evaluation myself. The hypothesis and the analysis are mine.

A citation identifies the source. Your AI-use disclosure explains how it contributed to your work and what thinking you did yourself. Neither style treats the AI itself as an author.

These are illustrative examples. Replace the dates, title, model and any details in braces with the real information for your work.
APA 7th edition Reference list — citing the tool generally
Anthropic. (2026). Claude [Large language model]. https://claude.ai
Use the tool’s latest update year, or its copyright year if no update date is available. Use 2026 only if it is the correct year. Do not turn the date you used the tool into a version number; include a version only when it is an actual, documented designation. Reference list — citing a specific, shareable conversation
Anthropic. (2026, September 4). Actual title of the conversation [Generative AI chat]. Claude. {shareable conversation URL}
Use the actual conversation date and title, and a link to that conversation. Include the actual model name if available. The conversation date is different from the update or copyright year used for a general tool reference. In your text — for either example, if the reference year is 2026
(Anthropic, 2026)
APA’s reference-format guidance (September 2025)
MLA 9th edition Works cited — title or description of the generated content
“Give me three counter-arguments to my conclusion that the experiment was reliable” prompt. Claude, model {actual model name}, Anthropic, 4 Sept. 2026, {shareable conversation URL}.
MLA omits the author element. Begin with a title or description of the generated content, which may include the prompt. Name the tool, its specific model, the developer as publisher, and the date the content was generated. Prefer a stable conversation link; use the general tool URL if a shareable link is unavailable. In your text — use a shortened title that clearly identifies the entry
(“Give me three counter-arguments”)
MLA’s updated citation guidance (August 2025)

Alongside any required citations, write a separate AI-use disclosure using the four cards above: the tool and model, the date you used it, your prompt, and how you used the output. Explain what you checked, changed or rejected and which thinking was your own. MLA: describing AI use beyond citation.

Where to go

The two lists of tools on this site, and what each one is approved for.

YCYW Tools

Agents, image, video and audio tools run by the Foundation. Approved for school work. Check each tool's label before you use it.

Browse YCYW Tools
External Tools

Models, services and mobile apps made by other companies. Some are approved and some are not. Check each tool's label before you use it.

Browse External Tools
Key terms

Eleven words used across the handbook and in class.

Thinker-in-chiefYou. You direct AI, judge what it gives you, and own the result.
Cognitive outsourcingDeciding on purpose to hand a specific, bounded sub-task to AI so you can focus on the thinking that matters.
Cognitive surrenderAccepting what AI says without thinking. The thing this framework exists to prevent.
Distributed cognitionThinking spread across you, your peers, teachers, sources, tools and AI. You direct it.
MetacognitionThinking about your own thinking. Knowing what you know, what you don't, and what needs your reasoning.
AI-embedded metacognitionUsing AI to help you learn how to think, instead of using it to produce answers for you.
AI² (AI-enhanced inquiry)Using AI as a thinking partner while you investigate something. You ask the question, judge the evidence, and defend your conclusion.
AI literacyFour domains: engage with AI, create with AI, manage AI, design AI. The entry point, not the goal.
HallucinationAI presenting invented information as fact, with confidence.
AI-enriched learningYou decided what to think, checked the output, can explain every part, and your reasoning got stronger.
AI-diminished learningYou skipped your own thinking, copied the output, and could not explain the work.
YCYW AI Handbook for Students and Teachers Full document, 17 pages. Also available in 中文.
Open the handbook