Young Lab for parents and teachers

Learners will be able to

• Tell the difference between a machine that follows fixed rules and one that learned from examples.
• Name three places they already meet AI in daily life.
• Explain why some things are "partly AI".

Talk about it

• Which answer surprised you? Why?
• What in our home do you think uses AI? How could we check?
• Is something better just because it uses AI?

Unplugged version

1. Walk around your home or classroom and list ten machines or apps.
2. Next to each, write: fixed rules, learned from examples, or not sure.
3. For each "learned" one, write what examples it might have learned from.

Learners will be able to

• Describe training and testing in their own words.
• Predict how a nearest-neighbour model will guess from its examples.
• Explain how missing examples create bias, and how to fix it.

Talk about it

• Why did the robot think the green apple wasn't an apple?
• If a computer only saw photos of some kinds of people, what might go wrong?
• Who should decide which examples a machine learns from?

Unplugged version

1. Draw a big square. Label left to right: green, yellow, red. Bottom to top: long, round.
2. Draw three red apples (A) and a lime, a lemon and a banana (N) where they belong.
3. Now place a green apple. Which letter is closest? That is the machine's guess. Is it right?
4. Add a green apple (A) as an example and try again.

Learners will be able to

• Spot common clues in made-up posts: no source, pressure to share, too good to be true.
• Spot common glitches in generated pictures: odd text, impossible shadows, extra parts.
• Use the question "Who made this, and why?" before sharing.

Talk about it

• Have you seen something online that turned out not to be true?
• Why might someone make a fake post?
• Who could you ask when you're not sure?

Unplugged version

1. Find two short posts or headlines (with a grown-up).
2. For each, answer: Who made it? Why? Where did the facts come from?
3. Circle any pressure words, like "share now" or "they don't want you to know".

Learners will be able to

• Name the parts of a clear prompt: role, task, detail, tone.
• Improve a vague request by adding specific details.
• Recognize personal information and keep it out of prompts.

Talk about it

• What's the difference between "make it good" and "in five sentences for a ten-year-old"?
• Why shouldn't we put our name, school or address in a prompt?
• When is it better to ask a person than an AI tool?

Unplugged version

1. Cut four strips of paper: role, task, detail, tone.
2. Write a choice on each and read the full prompt aloud.
3. Swap one strip at a time. Which change makes the request clearest?

Learners will be able to

• Judge whether an automated decision is fair and say why.
• Connect fairness to rights: being seen, being included, asking why, appealing to a person.
• Suggest a concrete way to make an unfair system fairer.

Talk about it

• Have you ever felt a rule was unfair? What made it unfair?
• Why does it matter that a person can check what a machine decided?
• How can we make sure technology includes everyone?

Unplugged version

1. Read each story card aloud.
2. Vote fair or unfair with a thumbs up or down.
3. For each unfair one, draw or write one fix.

Learners will be able to

• Explain next-word prediction using counts from training text.
• Describe how randomness (temperature) changes output.
• Explain why fluent text can still be wrong.

Talk about it

• Why can a chatbot sound confident and still be wrong?
• What would happen if the training text were unkind or one-sided?
• How is your own writing different from next-word prediction?

Unplugged version

1. Write five short sentences on a page. This is your training data.
2. Pick a start word. Count which word follows it most often. Write it down.
3. Repeat until you reach a full stop. Does the sentence make sense? Is it true?

Learners will be able to

• Identify personal and sensitive information.
• Choose safe settings for location, permissions and passwords.
• Respect other people's privacy and consent when sharing photos.

Talk about it

• Which apps on your phone know where you are? Do they need to?
• What would you never type into a chatbot?
• How do you ask a friend before posting their photo?

Unplugged version

1. List the apps you use most.
2. For each: does it use location, camera, microphone, contacts? Does it need to?
3. Turn off one permission you don't need, together with a grown-up.

Learners will be able to

• Design if-then rules and predict which rule will fire.
• Explain why rule order matters, and why safety rules go first.
• Compare a rule-based bot with an AI chatbot.

Talk about it

• When should a bot say "let me get a person"?
• What's one thing a rule-based bot does better than an AI chatbot? And worse?
• Who is responsible when a bot gives a bad answer?

Unplugged version

1. Draw a start box: "Message arrives".
2. Draw diamonds for questions like "Does it mention 'hours'?" with yes and no arrows.
3. Put a safety diamond first. Test your chart with five messages a friend writes.

Learners will be able to

• Name several AI-related careers, including non-technical ones.
• Connect school subjects and personal strengths to those careers.

Talk about it

• Which job surprised you most?
• What problem in our community would you like technology to help with?

Unplugged version

1. Pick one career card. Write what a day in that job might look like.
2. Find one person who does something similar and prepare three questions for them.