A Plain-English AI Glossary

The Words You'll Hear

The only AI words you actually need — in language that makes sense, with none of the jargon that makes good people feel left out.

The AI world is full of words that can make you feel like you don't belong in the conversation.
You belong in the conversation.
Free download

Want the printable version?

Pop in your email and we'll send you the designed PDF — keep it nearby, print it, share it. You'll also get the occasional plain-English note from us. Unsubscribe anytime.

No spam. Just the words you'll hear — and the occasional note.

No memorizing

There's no test, and no jargon you have to earn. Keep this nearby and look up a word when it trips you up.

Easy to deeper

Words run from everyday to deeper-end. Start at the top — the basics are plenty to begin.

Honest, not hype

Where it helps, we tell you how a thing feels, and what's still unknown.

Section One

The everyday words

The handful you'll hear most. If you only learn these, you're already in the conversation.

AI Artificial Intelligence
Software that is very good at guessing what comes next, based on patterns in everything it has read. It feels smart. It doesn't truly know.
Model
The specific AI "brain" you're using — like Claude, ChatGPT, or Gemini. Different models have different strengths, but the same basic idea.
Chatbot
An AI you talk to. It waits for you to ask, and then it answers. The kind of AI most people meet first.
Prompt
What you type to the AI — your question or instruction. The quiet secret: a great prompt matters far less than the AI actually knowing who you are.
Generative AI
AI that makes new things — words, pictures, voices, or video — instead of just sorting or searching. "Generative" simply means it creates.
LLM Large Language Model
The technical name for this kind of AI. It was trained on huge amounts of text to predict language. You'll see it everywhere — now you can ignore it.
Token
The small pieces AI reads text in — usually a word or part of a word. Think of beads on a string. AI reads, thinks, and is often priced by the token.
Training data
Everything the AI read while it was learning — like all the books, articles, and websites it studied. Its answers are shaped by what it saw.
Hallucination
When AI confidently makes something up. It sounds true and isn't. Always double-check anything that matters.
Bias
When AI leans one way because the material it learned from leaned that way. Like a mirror with a smudge — it reflects us, smudges and all.
Prompt engineering
The skill of asking AI clearly so you get useful answers. Helpful to learn — but who you are still matters more than how clever your wording is.
Context window
How much the AI can hold in mind at once. Like short-term memory: when it fills up, the earliest things start to slip away.
Section Two

Getting things done

Words for AI that doesn't just talk — it acts, connects, and helps you do real work.

Agent
An AI set up to do a task for you, not just answer a question. It acts on your behalf — like a helper you can hand a job to.
Agentic
A fancy word that just means "able to act like an agent" — taking steps, not only chatting.
Agentic workflow
When AI doesn't just answer but runs the steps of a task toward a goal — it plans, uses tools, acts. The catch: it can only run a workflow you can name. Spell out how your work actually moves, and "agentic" becomes possible. Skip that, and you're just automating a blur.
Automation
Setting things up to run on their own, without you doing each step by hand. Done well, it gives you hours back. Done badly, it just scales the rush.
Workflow
The steps a task takes from start to finish. Like a recipe — once it's written down, AI can help you follow it the same way every time.
API & Integrations
A way for two apps to talk to each other. Think of a door between two rooms, so your tools can pass things back and forth.
Multimodal
AI that handles more than words — it can also work with pictures, sound, and video. "Modes" just means kinds of input.
Deepfake
A fake image, video, or voice made by AI to look or sound real. Getting more convincing — a reason to slow down and check the source.
Image generator
An AI that makes pictures from a few words you type. Midjourney and DALL·E are two you'll hear named most.
Section Three

The names you'll hear

The tools and companies people mention. You don't need all of them — one is plenty to start.

ChatGPT
The chatbot from a company called OpenAI. It's the tool that brought AI to everyone's attention.
Claude
The AI from a company called Anthropic, known for thoughtful, careful writing and working with your own files.
Gemini
Google's AI, built into many of the Google tools you may already use.
Copilot
Microsoft's AI helper, tucked inside apps like Word and Excel to assist as you work.
OpenAI · Anthropic · Google
Three of the big companies building these models. When people argue about "which AI is best," they usually mean these.
Open source vs. closed
Open: anyone can see and use the model freely. Closed: it stays locked inside the company that made it. Both have trade-offs.
Section Four

The big-picture words

How AI works under the hood, and the questions no one has fully answered yet.

Machine learning
How AI "learns" from lots of examples instead of being given strict rules. Show it enough, and it picks up the pattern.
Neural network
The kind of software AI is built on, loosely modeled on the brain. It learns by quietly adjusting millions of tiny connections.
Deep learning
Machine learning with many layers stacked up. "Deep" refers to the layers — and it's what powers most of today's AI.
NLP Natural Language Processing
Teaching computers to work with everyday human language — the field that makes chatbots possible.
AGI Artificial General Intelligence
A hoped-for (and feared) future AI as capable as a person at almost anything. It doesn't exist yet — and no one truly knows if or when it will.
Alignment
Making sure an AI's goals match what people actually want. Still an open problem the whole field is working on.
AI safety
The work of making AI helpful and not harmful — fair, honest, and under human care. A big, active area of research.
Parameters
The countless little settings inside a model that shape its answers. More isn't always better — it's one number among many.
Section Five

The deeper-end words

You'll meet these later, in Stay in the Room. Unfamiliar — not difficult. No rush.

Project · Custom Instructions · Gem
The spot in Claude, ChatGPT, or Gemini where you paste your Harness, so the AI always starts out knowing you.
GitHub
A filing cabinet on the internet where you can keep your files so any AI can read the latest version. Not coding.
Repository repo
A folder that lives on GitHub. That's really all it is.
Terminal
The window where you type short commands to your computer. Home of the most powerful AI tools. Unfamiliar, not difficult.
Claude Code
A version of Claude that runs in the terminal, right on your own computer, with your own files.
MCP & connectors emerging
New plumbing that lets AI safely reach your other apps and files. Early days — this is a fast-moving space worth watching.
Section Six

Our words

The ideas at the heart of how we teach AI at Attune. This is where it becomes yours.

Harness our term
One page that holds your values and your standards, which any AI runs inside. AI reads it so it works from you, not from the average.
Attuned operating system our term
What it feels like when you apply the Harness: the AI working from who you actually are, in every room you’re in. The Harness is the page; this is the result.
Context
Everything the AI can "see" in the moment: what you've told it, and the files it can reach. More of the right context = more you in the answer.
Capacity Liberation our term
AI handling what it can handle, so you're freed to do what only you can do. Hours back — not just tasks ticked off.
AI-fluent · AI-native
Fluent: you're comfortable working with AI. Native: it's become second nature. Like a language — you get there step by step.
Attunement
Starting with you — your values and your voice — before you reach for any tool. The inner work always comes first.
Where this goes next

You don't need all of these today. The basics are plenty to begin.

This glossary is the doorway. When you're ready to make AI work from you — not the generic average — here's the path we walk with you.

FREE
Build your HarnessWrite the one page that teaches AI who you are. It saves in your browser.
Start
Read the AI Field Guide — freeOur free AI Field Guide, Attuned — meet AI on your own terms, and find your note before you build.
→
Stay in the RoomThe full self-paced course on Kajabi: name what’s in the way, write your values, and build the Harness your AI runs inside.

Start free with your Harness · Attune · You belong in the conversation.