Aidhya
Free preview lesson

Aidhya learning experience · 10 minutes

AI Detective

Can you trust an AI answer?

AI can be helpful, creative and surprisingly confident. It can also be wrong. Learn a simple three-check routine before you trust or share an answer.

Start the lesson

1Watch

Meet your first suspicious answer

Watch how an answer that sounds certain can fail a simple fact-check.

Read the video transcript

AI Detective

Can you trust an AI answer?

Have you ever asked AI a question and received an answer that sounded completely certain?

Remember one simple rule.

Confidence is not evidence.

A suspicious answer

Imagine an AI says, “The first computer was invented in two thousand and five.”

It sounds clear and confident.

But computers existed many decades earlier.

The answer sounds right, but it is wrong.

Why AI can make mistakes

Generative AI learns patterns from a lot of information.

It uses those patterns to create likely answers.

This helps AI explain, brainstorm, and create.

But AI can also invent details, mix up facts, or use old information.

Use three checks

First, pause

Ask yourself, “What is the exact claim?”

Look for names, dates, numbers, quotations, and statements about the real world.

Second, find the source

Who created this information, and would they know the answer?

For a school holiday, trust the official school notice over a random post or chatbot.

Third, compare

Find another trustworthy source and ask whether it agrees.

Asking the same AI again is not independent fact checking.

Important decisions need stronger evidence.

Your AI Detective routine

Use AI to explore, explain, and create.

Before you trust or share a factual answer, remember three words.

Pause. Check. Compare.

Now take the AI Detective quiz on Aidhya.

2Learn

Why “sounds right” is not the same as “is right”

💬

AI predicts an answer

A generative AI learns patterns from large amounts of information. When you ask a question, it builds a likely response. That makes it excellent for explaining ideas, brainstorming and drafting—but it may also invent a detail, mix up two facts or use old information.

🔎

You decide what to trust

Look for checkable claims: names, numbers, dates, quotations and statements about the real world. The more important the decision, the stronger your evidence should be. Asking the AI again is not an independent fact-check.

3Practise

The three-check routine

1Pause

What exact claim am I being asked to believe?

2Check

Who is the source, and would they know?

3Compare

Does another trustworthy source agree?

TRY ITAn AI says: “Your school will be closed next Monday.”

Pause: the claim is about Monday's school schedule. Check: the AI is not responsible for that schedule. Compare: look at the official school notice or ask the school—not another AI chatbot.

4Quiz

Earn your AI Detective badge

Answer all five questions. You can try again.

1Why can an AI give a confident answer that is wrong?
2Which part of an AI response most clearly needs fact-checking?
3What should an AI Detective do first?
4Where is the best place to confirm the date of your school's holiday?
5One reliable-looking website agrees with the AI. What should you do next?

Bring Aidhya to your school

One small lesson. A complete AI literacy journey.

Aidhya combines grade-appropriate learning experiences with a school Olympiad for Grades 2–10.