Context and Clarity: Giving AI Exactly What It Needs, Part 1

The big idea

Most teens type 5-word prompts. Better prompts include WHO you are, WHAT you are doing, and WHY. The more context, the more useful the answer.

Some examples

• Bad: 'Write me an essay.' Good: 'Help me brainstorm a 500-word essay for my 9th grade English class about whether kids should have phones in school. My teacher likes specific examples.'
• Bad: 'Help with math.' Good: 'I am in 8th grade Algebra 1. I need help understanding why we factor quadratic equations. Explain like I am new to this.'
• Bad: 'Tell me about birds.' Good: 'I am doing a science fair project on hummingbirds. Give me 5 surprising facts I can use as hooks.'
• Bad: 'Write a poem.' Good: 'Write a 4-line birthday poem for my grandma who loves gardens and is turning 80.'

Try it!

Take a one-line prompt you would normally use. Rewrite it with: who you are, what you are doing, who the audience is, what 'good' looks like. Compare the two AI answers.

AI Cannot Read Your Mind: Spell Out What You Want

The big idea

AI seems magical when it 'gets' what you want. Actually, it is just guessing based on common patterns. Be explicit about what you want — do not assume AI will figure it out.

Some examples

• Do not assume AI knows: your age, grade, location, situation. TELL it.
• Do not assume AI knows the format you want. SPECIFY: list, paragraph, song, table?
• Do not assume AI knows the length you want. SAY: 5 sentences? 1 paragraph? 500 words?
• Do not assume AI knows the tone. STATE: serious, funny, casual, professional?

Try it!

Tell AI What Level You Are At

The big idea

AI does not know if you are new to a topic or expert. Tell it. The explanation changes completely based on what you say.

Some examples

• 'Explain like I am brand new to this.'
• 'Explain like I have studied this for 3 years.'
• 'I know the basics, but explain the advanced stuff in detail.'
• 'Pretend you are tutoring me — I am 13, just learning Algebra.'

Try it!

Understanding "Tell AI What Level You Are At" in practice: Prompting is a skill: the more specific and structured your input, the more useful the output. Beginner? Advanced? AI calibrates explanations to your level — but only if you tell it — and knowing how to apply this gives you a concrete advantage.

• Apply calibration in your prompting workflow to get better results
• Apply level matching in your prompting workflow to get better results
• Apply context in your prompting workflow to get better results

1. Rewrite one of your best prompts using role + context + task + format
2. Ask an AI to critique your prompt and suggest improvements
3. Compare outputs from two models using the same prompt

AI Cannot Read Your Mind: Be Specific About What You Want

The big idea

AI feels magical when answers fit. But it is not magic — it is pattern matching. Be specific about what you want, not assumptive that AI will figure it out.

Some examples

• Vague: 'Help with my essay.' Specific: 'Help me write the intro for my 5-paragraph essay on whether kids should have phones in school.'
• Vague: 'Make it better.' Specific: 'Make this paragraph shorter (under 50 words) and more emotional.'
• Vague: 'Write a story.' Specific: 'Write a 200-word story about a dog who can talk, in the style of a children's book.'
• Vague: 'Help with code.' Specific: 'Fix this Python error: [paste]. Explain why it broke.'

Try it!

Prompt chaining: breaking big AI tasks into steps

The big idea

When you ask AI to 'write me a 5-page research paper on climate change with sources,' it tends to fall apart. But if you chain prompts — outline first, then sections, then citations — you get something usable.

Some examples

• Prompt 1: 'Give me a 5-section outline for a paper on X.'
• Prompt 2: 'Write section 1 from that outline, ~200 words.'
• Prompt 3: 'Add three real source ideas I should look up for section 1.'
• Each prompt builds on the last, like Lego bricks.

Try it!

Pick a project you're stuck on. Write down the 3 chained prompts you'd send AI instead of one mega-prompt. Try them in order.

When AI sounds 100% sure but is 100% wrong

The big idea

AI never says 'I think' or 'maybe' unless you ask it to. So even when it makes things up, it sounds like a textbook. Confidence is a writing style for AI, not a sign of truth.

Some examples

• AI confidently inventing a fake book title or author.
• AI giving the wrong year for an event in a smooth, sure tone.
• AI 'quoting' a study that doesn't exist, complete with fake stats.
• AI explaining a made-up word as if it's real.

Try it!

Ask AI for three quotes from a niche book or movie. Then verify each one online. Count how many AI got wrong while sounding confident.

Prompting AI to explain at different ages

The big idea

This trick is gold for studying. Ask AI to explain something four ways: like you're 5, 10, 15, and a PhD student. Comparing the four versions shows you what the concept actually IS.

Some examples

• Explain photosynthesis at four levels
• Explain JWT tokens at four levels
• Explain capitalism at four levels
• Explain heartbreak at four levels

Try it!

Pick a topic from class you sort-of get. Ask AI to explain it at age 5, 10, 15, and PhD. Notice what survives every version — that's the core.

AI and Context Stuffing: Give AI All the Background Up Front

The big idea

Context stuffing means loading AI with the full picture before asking the question. Most people ask first and add context after — that wastes turns and gets bland answers.

Some examples

• Start with: 'I'm a 16-year-old applying to art schools. My portfolio is illustration. My target is RISD.'
• Then ask the actual question after the background.
• Include constraints up front: time, budget, skill level, due date.
• Paste the rubric or assignment if there is one.

Try it!

Take a question you're about to ask AI. Write three sentences of context first. Then ask. Compare to your usual style.