The KAGAMI mark КАГАМИ
Children's Librarypublished by: East Accelerator Foundationhost: kagami.bgfree for everyone
kagami.bg/en/deca · children's library · lesson · machine-readable viewVERIFIED 2026-10-22 · AI-assisted: yes · reviewer: awaiting teacher and psychologist
IDENTITY
id
DB-IB-11 · Asking AI - and Checking
track
IB · Internet & Safety · lesson 11
level
Tinkerer · grades 5–7 · ages 10–13
duration
10–15 min lesson; optional 10 min shared unplugged activity
school link
School-like reading tasks; no verified curriculum alignment claimed. Design brief: DigComp 3.0 areas 1 and 5 and AI competence; mapping proposed.
publisher
East Accelerator Foundation · host: kagami.bg
access
free · no account; scripted simulation works offline and collects and stores nothing; shared parent-mediated question mechanism retained
trust
VERIFIED 2026-10-22 · AI-assisted: yes · reviewer: awaiting teacher and psychologist
versions
bg: /deca/DB-IB-11_pitam-izkustven-intelekt.html · en: /en/deca/DB-IB-11_pitam-izkustven-intelekt.html · parent guides: /deca/pomagalo/DB-IB-11_pitam-izkustven-intelekt.html and /en/deca/pomagalo/DB-IB-11_pitam-izkustven-intelekt.html
LEARNING OBJECTIVES
KEY CONCEPTS
STRUCTURE
ANSWERS
SAFETY
NEXT
SOURCES · LICENCE
IMPACT METADATA
path
6 · Social impact (proposed)
determinant
education (proposed)
sdg_target
4.4 (proposed)
who_criterion
accessibility (proposed)
impact_type
education · prevention (proposed)
TAGS
kidsgrades-5-7AI-literacychecking-sourcesmasterunpluggedbgen
VERIFIED · 22.10.2026 AI help: yes reviewer: awaiting teacher and psychologist

Asking AI - and Checking

Ask clearly, check a separate source and decide when to use AI. Fictional offline simulation for ages 10–13, with no saved data.

⏱ 10–15 min Tinkerer Internet & Safety · 11 grades 5–7
You will be able to ask clearly, check a claim in a separate source and decide when to use AI and when not to.

For parents: a one-page guide — dinner questions and checking with a book.

01Remember

Before this lesson: DB-IB-06 · How AI “Guesses” — and When It Is Wrong.

Can AI be wrong?

Yes. I check its answer before using it.

Is a confident answer enough?

No. I look for what another source says.

What if I am unsure?

I stop and ask a grown-up I trust.

02Story

An imaginary class is making a noticeboard for a reading club. A reading passage and a meeting notice lie on the table. One child suggests trying a fictional AI helper. The helper in this story is an example written by us, not a real service.

Their first question is “Tell me about the club.” The answer is long, but the board only needs the meeting place and day. The child changes the question: “Using the club notice, give the place and day in one sentence.” Now the task is clear.

The helper sounds certain. Someone starts to copy. Another child asks: “Which line in the notice confirms the day?” They open the sheet and compare that exact word. They find a mismatch. The confident tone did not help them spot it.

The class corrects the day using the notice. They use the checked information for the heading and choose their own idea for the decoration. They leave room for a rule: “Ask clearly, check separately, and decide for ourselves.” They may borrow an idea from a helper, but they decide what goes on their board.

The story and class are made up. In the next activity, you choose what to check and how.

03Worked example

  1. Ask clearly. Instead of “Help”: “Using this fictional passage, tell me where the club meets, in one sentence.” This prompt gives a task, context and an answer format.

  2. Separate out the claim. “The meeting is in the library.” I check the place, rather than how confident the sentence sounds.

  3. Compare. I read the line about the place in the separate passage. If I cannot find it, I ask a teacher to help. Another answer from the same helper is not a separate source.

  4. Decide. If the sources agree, I use the checked information in my own words. If they disagree, I correct it using the source; without evidence, I leave the question open. A clear prompt does not promise a correct answer.

04Try it now

This is a fictional simulation, not real AI. All answers are scripted. There are three school-like questions and one wrong answer. You only choose buttons; the simulation sends and saves nothing. It works without the internet.

1 · Predict

Can a clear question receive a wrong answer? Choose a prediction.

2 · Run

Choose a question, then reveal the answer.

3 · Change: choose your check

What will you check first? Then choose how. You can change your choices.

Reveal an answer first.

What will you do after checking?

05No screen

For about 10 minutes: choose a question you can look up in a book at home. Ask a family member the same question. They do not need to know the answer.

  1. Write the question and listen to your family member’s answer.

  2. Find a relevant passage in the book. Note the page and compare the exact claim.

  3. Say what agrees, what differs and what remains unclear. If answers differ, find another source with a grown-up; agreement alone is not proof.

06Make your own

“Our family rule for AI”: write one sentence on paper, in your own words — when we ask and when we do not.

An idea: “We ask AI for ideas, check claims in a book, and when we want a conversation with someone close, we talk to them.” Change it to fit your choice. Read your rule to a family member and explain why you chose it.

07What I learned

Check yourself

1. Which prompt is clearer?

2. How do you check the day?

3. When may you choose not to use AI?

Skill badge: “I ask and check”
It is yours when you can ask clearly, show a separate source and explain your decision. No points or rankings.

If something scares you, tell a grown-up you trust.

08I have a question

Do you still have a question? Write it here. It goes to your parent, not to strangers. Your parent confirms their e-mail first.

The form is loading… If it does not appear, write your question on paper and show it to a parent or teacher.

·Sources