Lesson: Step-by-Step Instructions — Teaching a Robot to Make a Sandwich

Field	Value
Subject	Computing
Domain	Artificial Intelligence
Age band	5–7 years
Type	Conceptual
Centrality	Low-medium (foundational for algorithmic thinking)
Taxonomy ID	mt__p5n8z5soJ
Standards	UK KS1 Computing (Algorithms) · CSTA K-1B (1B-AP-08, 1B-AP-11)
Tailored for	Gifted 5y9m, IQ 125-130+, asynchronous — verbal/reasoning well above age, emotionally 5

Read this first. Your son likely already understands "instructions" in the everyday sense — he can tell you how to brush teeth or get dressed. The lesson here is the conceptual leap: that computers cannot infer intent, cannot fill gaps, and execute literally. That idea is genuinely counterintuitive to a bright 5-year-old who reads social cues beautifully. The sandwich-robot game makes it physical and funny. Run the 60-second check at the bottom; if he sails through, treat this lesson as a 5-minute launchpad into the Stretch work, where the real thinking lives.

---

Why this matters

This is your son's first encounter with one of the foundational ideas in computer science: the algorithm — a precise, ordered, complete set of instructions. Before he can understand how voice assistants work, how game AI makes decisions, or how large language models generate text, he needs the bedrock insight that computers do only what they are told, exactly as they are told.

For a child like yours — verbal, reasoning-heavy, fast — this lesson is less about "what is an instruction" and more about confronting the gap between human communication (messy, context-rich, full of assumptions) and machine communication (literal, brittle, assumption-free). That gap is the source of nearly every software bug, every AI hallucination, every "Siri misunderstood me" moment. He will meet that gap again and again.

Some parents find this lesson lands fast and the child wants more. If that happens, you haven't done it wrong — you've arrived at the right place. Jump to Stretch.

---

Learning objective

Your son will understand that a set of instructions must be precise, ordered, and complete for a machine (which cannot guess intent) to execute it correctly, and he will be able to demonstrate this by giving and debugging instructions for a simple physical task.

You want to hear him say something like:

"The robot can't know I meant 'use the knife' unless I say it. It just does exactly what I tell it."

---

Before you sit down together

Materials

• A loaf of bread (or two slices on a plate) — real food makes it memorable
• Butter or jam in a jar with a lid — the lid matters (see below)
• A butter knife — keep it child-safe
• A plate
• Paper and a pencil/crayon — for writing or drawing the steps if he wants to
• A small toy robot or stuffed animal — optional, to be "the robot's brain" you hand instructions to

Why these and not a worksheet: the physical comedy of watching you (the robot) misinterpret his instructions is the entire pedagogical engine. A worksheet abstracts away the thing he needs to feel — the frustration and hilarity of literalism.

Best time of day for this lesson

Mid-morning, after a snack and some movement, tends to work well for this age band — your son is fed, alert, and in a playful mode. Avoid the post-lunch slump and the pre-dinner cranky window. This lesson benefits from silliness, so don't attempt it when either of you is tired or rushed.

Some parents find that doing this as a "kitchen game" rather than a "lesson" entirely changes the tone — if your son resists anything that smells like school, try framing it as "I'm going to pretend to be a robot and you have to program me."

---

Activity: "The Jam-Sandwich Robot"

A four-phase conceptual lesson (Introduce → Explore → Apply → Wrap-up). Total: 15–20 minutes.

Parent-as-robot rule: Once you announce you are the robot, you follow instructions exactly and literally. If he says "put the jam on the bread," you place the entire closed jar on top of the bread. If he says "use the knife," you hold the knife but do nothing else until told. Do not hint. Do not fill gaps. The comedy IS the lesson.

---

Phase 1 — Introduce (3–4 min)

Sit at the table with materials in front of you. Keep it light.

Sample dialogue:

"I'm going to be a robot today. Robots are really good at following instructions, but they can't think for themselves. They don't know what you 'mean' — they only know what you 'say.' So if you want me to make a jam sandwich, you have to tell me every single step, in order, and I'll do EXACTLY what you tell me. Not what you meant. What you SAID. Ready?"

You might let him try one practice command to warm up:

• "Robot, pick up the bread" → you pick up the whole loaf.
• "No! Just one slice!" → you have just taught him the lesson yourself. Smile.

This phase should feel like a game reveal, not a lecture.

---

Phase 2 — Explore (6–8 min)

Now he programs you to make the sandwich. You follow his instructions literally and let the comedy emerge.

Let him dictate. Don't correct ahead of time — let the mistake happen, then freeze and ask:

• "I'm holding the jar of jam on top of the bread. Is that what you wanted? That's what you told me to do."

Common sequences you might live through:

• "Put the jam on the bread" → jar, lid on, placed on bread.
• "Take the lid off!" → lid off, jar still on bread.
• "Spread the jam!" → you spread jam on the outside of the jar, or on the lid.
• "Open the bread!" → you tear the bag open wildly.

After each misfire, pause and ask:

• "What did you MEAN? What instruction would have told me to do that?"

This is the debugging instinct. You are teaching him to revise instructions, not to blame the robot. Some gifted children find this hilarious; a few find it frustrating. If frustration appears, see the kid-response scripts below.

Rich language to drop naturally during this phase: sequence, precise, missing step, literal, debug.

---

Phase 3 — Apply (4–5 min)

Now flip it. Have HIM be the robot, and YOU give the instructions. But leave a deliberate gap and see if he notices.

Sample dialogue:

"Okay, now you're the robot. I'm going to give you instructions. Robot: open the bread bag. Robot: take out one slice. Robot: put jam on the bread."

Pause. Did he ask how much? Did he ask which side? If he just does it, celebrate and then reflect:

• "Wait — how did you know how much jam I wanted? You guessed! But a real robot can't guess. So what instruction was I missing?"

This phase is where the concept clicks — he experiences from the inside how natural it is to infer, and that helps him understand why machines need extra help.

---

Phase 4 — Wrap-up (2–3 min)

Put the sandwich aside (or eat it — that's fine) and sit together for a brief reflection. You might ask:

• "What was the hardest part about giving me instructions?"
• "What's one instruction you had to fix?"
• "Why can't a robot just guess what you mean?"

You're looking for him to articulate the core insight in his own words. Don't supply the answer — give him space. If he says something like "because robots are dumb," gently reframe: "Not dumb — they just do exactly what you tell them. So we have to be really careful with our words."

Optionally: have him draw or write 3–5 steps for brushing his teeth or making his bed, using precise language. Some gifted 5-year-olds love this; others find writing tedious. Follow his lead.

---

Kid-response scripts

He says…	What's happening	You might try…
"This is so silly!" (laughing)	He's fully engaged and feeling the comedy — ideal state	Ride the wave. Say: "It IS silly. And this is exactly why programmers laugh at their computers sometimes." Keep going.
"No, that's not what I meant!" (frustrated)	He's conflating "what I meant" with "what I said" — the exact gap	Pause gently: "You're right — that wasn't what you meant. But it WAS what you SAID. What could you say differently?"
"Just do it normally!"	He wants to skip to the outcome without specifying process	"A real robot has no idea what 'normally' means. Can you teach me what 'normally' looks like, step by step?"
"Robots aren't really like this!"	He's pushing back — possibly knows more than expected	Excellent. Ask: "You're right that some robots are smarter now. But how do you think they GOT smarter?" This opens the door to Stretch.
"I already know how to make a sandwich."	Procedural mastery is high; he thinks the lesson is beneath him	"You do! But can you teach someone who's NEVER seen a sandwich before? That's different. Try me."
"What if I just show you?"	He wants to demonstrate rather than verbalize — strong kinesthetic learner	"Great — but a robot across the world can't see you. How would you send your instructions in words?"
(silence, thinking hard)	He's debugging internally — do not interrupt	Wait. Count to ten silently. Some of the best learning happens here.

---

Common misconceptions to watch for

What you see	What's actually going on	How to gently address
He gives instructions that assume knowledge ("spread it") without specifying tool or motion	He hasn't yet separated his own knowledge from the robot's (zero) knowledge — this is called the curse of knowledge	Narrate the gap: "I've never seen a sandwich before. What does 'spread' mean? What do I use?"
He lists steps out of order ("Eat it… put the jam on… get the bread")	He's focused on content, not sequence — he knows the goal but not the order matters	Act it out: eat imaginary bread, then try to add jam. "Hmm. What went wrong with the order?"
He thinks the robot "should know" obvious things	He's anthropomorphizing — projecting human common sense onto a machine	"You're so right that a PERSON would know that. But a machine doesn't have a brain like ours. It only has the words we give it."
He memorizes "robots do what you say" as a fact without really feeling the literalism	Classic gifted pattern — procedure without concept	Don't accept the verbal answer. Return to the physical game and let him experience a misfire again.
He gets upset when you "do it wrong"	He may be reading your literalism as you being unkind or deliberately obtuse	Reassure: "I'm not teasing you. This is really how computers work. Programmers get frustrated too — let me show you something funny about that…"

---

Stretch (where the real lesson lives for your son)

If your son finishes the core activity and says "that was easy" or finishes your sentences, these extensions are where the actual learning happens for a child at his level. Pick one or two, not all.

Stretch 1 — "What if a step is missing?" (debugging, ~5 min)

Write down a list of 5–7 steps for making the sandwich, but deliberately remove one step (e.g., leave out "take the lid off the jar"). Hand him the list and ask:

"A robot was given these instructions and the sandwich went wrong. Can you find the missing step?"

This introduces debugging as a named concept and asks him to reason backwards from effect to cause — a high-level cognitive move he's likely ready for.

Stretch 2 — Loops and repeats (~5 min)

Ask: "What if I told you: 'Take one step. Take one step. Take one step…' for the whole hallway? Is there a shorter way to say that?"

If he says "just say walk down the hall" — celebrate, then introduce the word repeat or loop. You might try:

"Programmers use a shortcut called a LOOP. You say: 'Repeat: take one step — until you reach the door.' Can you write a loop for brushing your teeth? How many times should the brush go back and forth?"

This plants the seed for iteration, a concept he'll meet formally in any programming environment.

Stretch 3 — Conditional instructions (if/then, ~5 min)

"What if the robot is making tea, and the water is already hot? Should it still boil the water again? How would you tell the robot to decide?"

If he's intrigued, introduce: "IF the water is cold, THEN boil it. IF the water is hot, THEN skip to pouring."

Some 5-year-olds with this profile find conditionals absolutely delightful. If he does, you've opened the door to decision-making in AI, which connects directly to the dependent topics in this strand.

Stretch 4 — Ambiguity detectives (~5 min)

Give him two instructions and ask which is "robot-safe":

• "Put the jam on the bread."
• "Use the knife to spread one spoonful of jam evenly on one side of one slice of bread."

Ask: "Which one would a robot understand? Why? What's wrong with the first one?"

This trains his ear for precision and ambiguity — skills that underpin all future work in writing, science, and computing.

Stretch 5 — Teach the robot to teach a robot (~5–7 min, if he's flying)

"What if you had to write instructions so that a robot could teach ANOTHER robot how to make a sandwich? Now you're writing instructions about instructions."

This is metacognition meets recursion. He probably won't solve it fully, but the question itself will make his brain buzz. Let him think out loud. Don't rush to resolve it.

---

Quick mastery check (60 seconds)

Ask these three prompts in sequence. Check each box only if he answers in a way that shows genuine understanding (not just echoing the right word).

• [ ] "Can you tell me one rule about how robots follow instructions?" (Looking for: they do exactly what you say / they can't guess / they need every step)
• [ ] "What would happen if I told a robot 'make a sandwich' and nothing else?" (Looking for recognition that the instruction is incomplete)
• [ ] "Why do instructions need to be in order?" (Looking for: because the robot would do things at the wrong time / it would go wrong)

If he checks all three confidently — skip to Stretch immediately. The core lesson is already his.

---

Formal mastery check

From the taxonomy evidence field, your son should be able to:

1. Write or describe step-by-step instructions for a simple everyday task (e.g., making a jam sandwich, brushing teeth, getting dressed)
2. Predict what goes wrong if a step is missing or in the wrong order — you can test this by removing a step from a written list and asking him to find the bug
3. Explain that a computer follows instructions exactly — it cannot guess what you meant — this should come out in his own words, not as a memorized phrase

Assessment prompt from the dataset:

If {{name}} were to write instructions for making a jam sandwich so precise that even a robot could follow them, could they include every single step?

Watch for whether he naturally includes "invisible" steps like opening the jar, picking up the knife, or specifying which side of the bread. Those are the markers of real understanding.

---

Vocabulary to use naturally

Drop these into conversation without making a big deal of them. Your son will absorb them through context:

• Instruction — a single command telling someone (or something) what to do
• Sequence — the order in which instructions happen
• Precise — exact, with no ambiguity
• Literal — doing exactly what was said, not what was meant
• Debug — finding and fixing a mistake in instructions
• Algorithm — a complete set of steps to solve a problem (you might save this word for the wrap-up, or let it appear naturally)

---

What comes next

This lesson is a prerequisite for several topics in the computing strand. Once your son has internalized that machines follow instructions literally, he's ready for:

• Data and information in computers (hard dependency) — he can now grasp that data is just another form of precise input that computers process according to instructions.
• Voice assistants and how they work (soft dependency) — understanding algorithms helps him see why Siri or Alexa sometimes mishear him: the speech-to-instruction pipeline has gaps.
• AI in computer games (hard dependency) — before he can understand how game enemies "decide" to move, he needs to understand that those decisions are instructions executing, not minds thinking.

You might find that voice assistants are the most motivating next step for a 5-year-old — they're part of his daily life and the "robot gets it wrong" experience maps directly onto "Siri didn't understand me."

---

If this lesson didn't land

Some days, even the best lesson flops. That's information, not failure. Consider:

• Try a different physical task. If the sandwich didn't spark, try "program me to put on my shoes" or "program me to walk to the front door." The principle is the same; the context may resonate better.
• Shift the time of day. If he was tired or hungry, retry after a snack or the next morning. This lesson depends on playful energy.
• Shorten dramatically. Drop the wrap-up and just do the sandwich robot for 5 minutes. End on a laugh. Come back to reflection tomorrow.
• Let him be the robot first. Some children resist being the instructor but love being the literal one. Flip the roles and let him experience the comedy from your side.
• Check the prerequisite. If he didn't seem to understand what a robot or computer is in the first place, back up to "Computers in Everyday Life" before returning here. You can't teach literalism to a child who doesn't yet have a mental model of what a computer is.

---

Source

• Taxonomy ID: mt__p5n8z5soJ
• Dataset: Primary Computing / AI strand (ages 5–7)
• Standards: UK KS1 Computing (Algorithms) · CSTA K-12 (1B-AP-08: Compare and refine multiple algorithms for the same task; 1B-AP-11: Decompose problems into smaller, manageable subproblems)
• Generated by: Parent-facing lesson planner, tailored for gifted asynchronous 5y9m (IQ 125-130+)