THE AI GLOSSARY FOR STUDENTS

Software that learns how to do something by finding patterns in huge amounts of examples — instead of following rules a person wrote out.

In practice — A chatbot answering a question nobody ever wrote a rule for.

Taught in Lesson 0 · Welcome to AI Explorers — watch it free · More on Artificial Intelligence

Something that repeats often enough that AI can spot it and use it to make a guess about what comes next.

In practice — After seeing millions of sentences, AI learns that "peanut butter and" is usually followed by "jelly."

Taught in Lesson 0 · Welcome to AI Explorers — watch it free · More on Pattern

The pile of examples an AI learns from. Swap that pile and the AI behaves differently — even if nothing else about it changes.

In practice — An AI trained only on cat photos will not recognize a dog.

Taught in Lesson 0 · Welcome to AI Explorers — watch it free · More on Training Data

The trained "brain" of an AI — what is left after it has learned all its patterns. Two of them can behave differently if they learned from different examples.

In practice — ChatGPT, Claude, and Gemini are different models.

Taught in Lesson 0 · Welcome to AI Explorers — watch it free · More on Model

A set of steps written by a person to solve a problem. Steps like these can be clever without being AI — a calculator follows them.

In practice — Shortest-route math in a GPS is an algorithm, not AI.

Taught in Lesson 0 · Welcome to AI Explorers — watch it free · More on Algorithm

AI's best guess at what comes next, based on patterns. Your feed, your recommendations, and autocomplete all work this way — no editor is choosing for you.

In practice — When a chatbot writes, it is choosing the next likely word.

Taught in Lesson 1 · AI vs. Humans: What Makes Us Different? · More on Prediction

Deciding what actually matters, what is fair, and what is right. AI can be fast and patterned; this is the part only a person can do.

In practice — Deciding which teammate gets the last roster spot.

Taught in Lesson 1 · AI vs. Humans: What Makes Us Different? · More on Human Judgment

Deciding which jobs go to AI, which stay human, and which are a team-up. It is the first of the 4 D's.

In practice — Letting AI draft rhymes, but you pick the one that sounds like Grandma.

Taught in Lesson 1 · AI vs. Humans: What Makes Us Different? · More on Delegation

What you type to an AI. Your exact words change what comes back.

In practice — "Act as a zookeeper and describe red pandas in 4 bullets."

Taught in Lesson 2 · Your Words Are the Superpower · More on Prompt

Being exact instead of vague. It shrinks the AI's choices down to the answer you actually wanted — vague in, vague out.

In practice — "4 bullet points" instead of "make it good."

Taught in Lesson 2 · Your Words Are the Superpower · More on Specificity

The background information you give the AI so it does not have to guess — who you are, what you already tried, what you need it for.

In practice — "My science fair is Friday and I have only finished the poster."

Taught in Lesson 2 · Your Words Are the Superpower · More on Context

Communicating clearly with AI — the second of the 4 D's. It covers what you want, how you want it made, and how good it has to be.

In practice — Naming a role, a task, and a format in one prompt.

Taught in Lesson 2 · Your Words Are the Superpower · More on Description

A reusable structure for asking, instead of guessing each time. A strong one uses the 4 R's to shrink the AI's choices.

In practice — Iara's Meta Prompt Card, reused for every new question.

Taught in Lesson 3 · The 4 R's: Your Meta Prompt Framework · More on Meta-Prompt

Improving your prompt based on what came back, then running it again. Your first prompt is only a draft — the loop is the real skill.

In practice — Round 1 too long → add "in 3 bullets" → run again.

Taught in Lesson 4 · Level Up: Iteration & Chain-of-Thought · More on Iteration

Asking the AI to work through its reasoning step by step, which makes it noticeably better at hard problems.

In practice — "Show your steps before you give the final answer."

Taught in Lesson 4 · Level Up: Iteration & Chain-of-Thought · More on Chain-of-Thought

Judging whether an AI output is actually any good — accurate, useful, and well reasoned. The third of the 4 D's.

In practice — Picking the better of two answers and saying WHY.

Taught in Lesson 4 · Level Up: Iteration & Chain-of-Thought · More on Discernment

AI turning words into pictures by mixing patterns from millions of examples. Vague words let the AI decide.

In practice — Typing a description and getting an original illustration back.

Taught in Lesson 5 · Painting with Words: AI Image Creation · More on Image Generation

What is actually in the picture — the first of the five image ingredients.

In practice — "A fluffy orange cat."

Taught in Lesson 5 · Painting with Words: AI Image Creation · More on Subject

How the picture is made — watercolor, photo, pixel art, 3D render. Leave it out and the AI picks for you.

In practice — "Watercolor illustration" vs "photorealistic."

Taught in Lesson 5 · Painting with Words: AI Image Creation · More on Style

The feeling the picture gives off — cozy, tense, lonely, joyful.

In practice — Same lighthouse: "calm" vs "stormy" are different pictures.

Taught in Lesson 5 · Painting with Words: AI Image Creation · More on Mood

Where the light comes from and what it does to the scene — golden sunset, neon night, single spotlight.

In practice — "Lit only by a computer screen."

Taught in Lesson 5 · Painting with Words: AI Image Creation · More on Lighting

Where the viewer is standing — close-up, wide shot, overhead, or looking up from below.

In practice — "Close-up at eye level" vs "overhead shot."

Taught in Lesson 5 · Painting with Words: AI Image Creation · More on Camera Angle

Being the deliberate director of style, mood, lighting, and framing — instead of accepting whatever the AI produced first.

In practice — Rejecting a beautiful image because the brief said watercolor.

Taught in Lesson 6 · Style, Mood & Iteration: AI Art Director · More on Art Direction

Telling the AI what you do NOT want in the picture.

In practice — "No people, no text, no watermarks."

Taught in Lesson 6 · Style, Mood & Iteration: AI Art Director · More on Negative Prompt

What you include in the shot and what you deliberately leave out.

In practice — Cropping tight on the mascot instead of showing the whole room.

Taught in Lesson 6 · Style, Mood & Iteration: AI Art Director · More on Framing

A shot-by-shot sketch of your video before anything gets generated. Directors decide, then delegate.

In practice — Three stick-figure panels for a 10-second film.

Taught in Lesson 7 · Lights, Camera, AI: Directing Video & Audio · More on Storyboard

AI-made video or audio of a real person doing or saying something they never did. Making one without permission is a severe ethical violation.

In practice — Making a video of a classmate saying something they never said.

Taught in Lesson 7 · Lights, Camera, AI: Directing Video & Audio · More on Deepfake

A real person's face, body, or voice. You need their explicit permission before using it — in this program we use fictional characters only.

In practice — Why every program character is invented, never a real classmate.

Taught in Lesson 7 · Lights, Camera, AI: Directing Video & Audio · More on Likeness

Clear permission from a person before you use their likeness or voice. Silence does not count.

In practice — A signed consent file before any real likeness is generated.

Taught in Lesson 7 · Lights, Camera, AI: Directing Video & Audio · More on Consent

An AI with standing instructions — a system prompt it reads before every single chat, so it behaves the same way every time.

In practice — A study helper that always hints instead of answering.

Taught in Lesson 8 · Design Your AI: What Should It Do? · More on AI Assistant

The plan you write BEFORE building an assistant, covering every layer it will run on. Plan first, then climb.

In practice — Purpose, personality, and guardrail, decided on paper first.

Taught in Lesson 8 · Design Your AI: What Should It Do? · More on Blueprint

The standing instructions an assistant reads before every chat. Students build theirs layer by layer, deciding exactly how their assistant should behave.

In practice — The text that turns a blank AI into YOUR tutor.

Taught in Lesson 9 · Build It: Your Custom AI Assistant · More on System Prompt

The hard limit your assistant must not cross. It is the most important guardrail — it decides whether your assistant is safe.

In practice — "Never ask for a real name, school, or address."

Taught in Lesson 9 · Build It: Your Custom AI Assistant · More on Never-Do Rule

Finding which part of a system caused a problem and fixing that part — not rewriting everything.

In practice — Wrong facts → fix the Knowledge layer, not the Tone layer.

Taught in Lesson 9 · Build It: Your Custom AI Assistant · More on Debugging

Something that runs by itself: IF a trigger happens, THEN an action follows — with no human in between the steps.

In practice — Photos backing up every night at 2am.

Taught in Lesson 10 · AI on Autopilot: Triggers, Actions & Automation · More on Automation

The event that starts an automation — the "IF" half.

In practice — "Every morning at 7am" or "when a new email arrives."

Taught in Lesson 10 · AI on Autopilot: Triggers, Actions & Automation · More on Trigger

What the automation actually does once it fires — the "THEN" half.

In practice — "Send the family a plan-B message."

Taught in Lesson 10 · AI on Autopilot: Triggers, Actions & Automation · More on Action

An AI given a goal that plans its own steps to get there. It does not run in a straight line — it navigates decisions.

In practice — An agent told "plan a beach day" that checks weather first.

Taught in Lesson 11 · AI Agents: Teaching AI to Plan Ahead · More on AI Agent

The branching map of choices an agent works through — if this, go here; if that, go there.

In practice — Rain over 50%? → plan B. Otherwise → beach.

Taught in Lesson 11 · AI Agents: Teaching AI to Plan Ahead · More on Decision Tree

A deliberate stop where the agent pauses and brings a person back in. An agent with a goal but no rules orders 400 pizzas.

In practice — Requiring a human OK before anything gets sent or spent.

Taught in Lesson 11 · AI Agents: Teaching AI to Plan Ahead · More on Human Checkpoint

Building a working mini-app by describing it in plain language instead of writing code — all four D's in action.

In practice — Describing a chore tracker precisely enough to be built.

Taught in Lesson 12 · Vibe Coding: From Idea to App · More on Vibe Coding

A description precise enough to build from. "Make it nice" is not one; "3 checkboxes per person" is.

In practice — Naming the exact screen, numbers, and what happens on tap.

Taught in Lesson 12 · Vibe Coding: From Idea to App · More on Spec

When AI states something false with total confidence. Its mistakes sound exactly as sure as its correct answers.

In practice — Inventing a report from an institute that does not exist.

Taught in Lesson 13 · Trust But Verify: AI Hallucinations & Fact-Checking · More on Hallucination

How sure an answer sounds tells you nothing about whether it is true. AI writes the most likely-sounding answer.

In practice — A made-up law quoted with a year and a section number.

Taught in Lesson 13 · Trust But Verify: AI Hallucinations & Fact-Checking · More on Confidence ≠ Evidence

Soleil's three steps for checking a claim: Flag it, Find a real source, Compare. You cannot use AI to fact-check AI.

In practice — Flagging "40-mile boardwalk," then checking the town website.

Taught in Lesson 13 · Trust But Verify: AI Hallucinations & Fact-Checking · More on Verification Move

Where a fact actually comes from — something you can go look at yourself. Real ones can be found; fake ones can only be repeated.

In practice — A museum's official site instead of "a 2019 report."

Taught in Lesson 13 · Trust But Verify: AI Hallucinations & Fact-Checking · More on Source

When an AI treats some people or groups unfairly because its examples were unbalanced. Always ask: whose data taught this AI?

In practice — A tool that understands some accents far better than others.

Taught in Lesson 14 · Fair, Safe & Honest: The Ethics of AI Use · More on Bias

Seo's pause before you share anything: Is it fair? Is it safe? Is it honest?

In practice — Running all three questions before posting a project.

Taught in Lesson 14 · Fair, Safe & Honest: The Ethics of AI Use · More on 3-Question Ethics Check

Owning what you publish. The choice to use AI carries responsibility — what you ship is yours.

In practice — "The AI made it" is not an excuse for what you handed in.

Taught in Lesson 14 · Fair, Safe & Honest: The Ethics of AI Use · More on Accountability

Using AI responsibly — ethics, transparency, and accountability. The fourth of the 4 D's.

In practice — Disclosing AI's role before anyone has to ask.

Taught in Lesson 14 · Fair, Safe & Honest: The Ethics of AI Use · More on Diligence

Using AI to make you think harder instead of thinking for you. The best use of AI is not getting answers — it is getting better questions.

In practice — Asking AI to quiz you instead of asking it to write it.

Taught in Lesson 15 · AI as Your Learning Partner · More on Learning Partner

Audrey's three-step self-check for making sure you actually learned something, not just got an answer.

In practice — Writing your own answer before you ever open the AI.

Taught in Lesson 15 · AI as Your Learning Partner · More on The 3 Checks

The note that says what AI did, what you did, and how you verified it. It goes on Slide 1 of your showcase and you read it aloud.

In practice — "AI drafted my prompts; I chose the concept and checked the facts."

Taught in Lesson 15 · AI as Your Learning Partner · More on Disclosure Statement

Doing your own thinking and being honest about any help you had — including AI.

In practice — Being able to explain every part of what you turned in.

Taught in Lesson 15 · AI as Your Learning Partner · More on Academic Integrity