The words

Plain-English AI glossary.

Every term you're likely to hear, defined in a sentence — then in a paragraph if you want the longer version. No jargon for jargon's sake. No pretending simple ideas are complicated.

AI Model a.k.a. "the model"

A trained pattern-predictor. Not a mind, not a database — a big set of numbers that learned how language works.

When people say "ChatGPT" or "Claude" or "Gemini," they usually mean a specific AI model plus the app that lets you talk to it. The model itself is just the trained system — the numbers — that takes your text in and produces text out.

Appears throughout. See also: Weights, LLM.

Prompt what you type

The question or instruction you give AI. The quality of your prompt drives the quality of the answer more than anything else.

"Help me with my lease" is a prompt. So is a three-paragraph briefing with goals, context, and expectations. The difference between those two is where the single biggest skill improvement comes from — see The Three Questions.

Featured at Station 4 & Station 7.

RTF Framework Role · Task · Format

A simple prompting template: tell AI who to be (Role), what to do (Task), and how to output (Format).

Example: "Role: Act as a property manager. Task: Summarize these three tenant emails. Format: One paragraph each, flag urgency." RTF is the fastest framework to remember and works across any AI tool. For Copilot-specific prompts, see GCES, which adds a Source field.

Taught in the full-length professional workshop.

GCES Framework Goal · Context · Expectations · Source

Microsoft's four-part prompting framework. Same idea as RTF, but adds Source — which is what turns Copilot from a chatbot into a co-worker who's read your files.

Goal — what outcome do you want?
Context — what background matters?
Expectations — what specific steps, format, length?
Source — which files or documents should it use?

Source is the most important part for anyone using Copilot with real documents. The Three Questions from the journey (What do I want? What would correct look like? What can't it know?) are the consumer-friendly cousin of GCES — same instincts, fewer letters.

Training

The one-time, expensive process of teaching an AI model how language works by feeding it billions of examples.

Training a frontier model like GPT-5 reportedly costs around $500 million in compute. Once training is done, the model is fixed — it doesn't learn from you when you chat with it. (It may learn in aggregate if your chats are used to train future versions — which is why privacy settings matter.)

Training Data

The text the model was trained on. Books, web pages, forums, documents. Tens of terabytes — large, but tiny compared to the internet.

AI models don't store this data. They extract patterns from it into their weights. That's why a 1–2 TB model can answer about almost any topic — it learned the patterns in a bigger body of text, not the text itself.

Training Cutoff

The date after which the model doesn't "know" what happened — unless it can search the web.

Every model has a point where its training data ends. After that date, the model is frozen — it literally hasn't seen anything newer. Modern models bridge this with web search, but the underlying training is still a snapshot.

See also: Web-enabled AI.

Weights / Parameters "the numbers"

The billions of numbers inside a model that encode what it learned. Frontier models have a trillion-plus weights.

When AI processes your prompt, it's doing math with these weights to predict the next word. All the "knowledge" the model has is compressed into these numbers — which is why a ~1–2 TB model can answer about almost anything, but also why it occasionally makes things up.

Inference running the model

When you use an AI to answer a question. As opposed to training it. Inference is cheap; training is expensive.

Every chat you have is an inference. Modern inference costs fractions of a cent per query — in part because the training already paid for the hard work. The cost-per-query has dropped roughly an order of magnitude in the last two years.

Token

The unit AI works in. Roughly a word, or a piece of a word. "Hello world" is about 2 tokens.

Models don't think in letters or sentences — they think in tokens. API pricing is usually per million tokens. Token limits (e.g., "128K context") tell you how much text a model can hold in its head at once.

See also: Context Window.

Context Window "what it can see"

How much text the model can hold in its head at once — including your messages, its replies, and any documents you pasted.

Modern models have context windows of 100,000 to over a million tokens. If a conversation goes longer than the context window, the earliest parts fall out — the model literally forgets them. Long chats drift; when that happens, start a new chat.

Deterministic vs. Probabilistic

A calculator is deterministic — same input, same output. AI is probabilistic — same question, different answer each time.

AI samples from probabilities each time it picks the next word. That's why "ask again" often gets a better answer, why you can't reliably reproduce a bad one, and why two people asking the same question get different responses. You saw this live at Station 3 of the journey.

Hallucination

When AI confidently invents a fact, citation, court case, quote, or statistic that isn't real.

Hallucinations come from the same engine that produces good answers. Models are trained to sound confident; they're not trained to say "I don't know." Newer models are measurably better at flagging uncertainty, but the core issue hasn't gone away — so verifying anything that matters remains the safe habit.

Track current rates: Vectara Hallucination Leaderboard.

Grounding

Pointing AI at your actual documents so it answers from them — not from memory.

The single biggest quality upgrade most people can make: instead of asking a question in the abstract, paste or attach the document and ask about that. Accuracy jumps from "sometimes" to "almost always." This is also the concept behind Copilot Notebooks, custom GPTs, and enterprise AI tools that plug into your files.

RAG Retrieval-Augmented Generation

The technique where AI first searches for relevant documents, then writes its answer using them. It's how "grounded" AI works under the hood.

When you upload a document to Copilot or ChatGPT and ask about it, RAG is what's happening: the system finds the relevant chunks, feeds them to the model as context, and the model answers from them. Why this matters: grounded answers hallucinate far less than ungrounded ones. Copilot Notebooks are essentially "a bucket of documents for RAG to use whenever you chat."

Web-enabled AI "AI with search"

AI that can search the internet in real time. Most modern chat tools (ChatGPT, Claude, Gemini, Copilot) do this automatically when they think it'll help.

This closes much of the gap from the training cutoff — the model can fetch current info and cite it. But the core engine is still pattern-based. It can grab stale articles, misread live status, or miss the most current source. Treat web-enabled AI as "strong first pass," not "verified answer."

Memory

An optional feature that lets AI remember things from past conversations. Not the same as "knowing you."

ChatGPT, Claude, and Gemini all let you turn on persistent memory. The AI stores facts you've told it (preferences, your job, your projects) and recalls them across chats. Useful — but it's opt-in, the model isn't "intelligent about you," and you should review what's been stored from time to time.

Agent

AI that takes multi-step actions, not just answers questions. "Review these files, summarize each, and email a report" — and it just does it.

Regular chat is one-and-done: you ask, it answers. An agent has a goal and works through steps — reading files, running tools, checking its work, asking clarifying questions. This is the frontier — powerful, but still needs human oversight for anything high-stakes.

LLM Large Language Model

The technology underneath ChatGPT, Claude, Gemini, and Copilot. "Large" because of how many weights and how much training data.

An LLM is the model itself — the trained pattern-predictor. Products like ChatGPT wrap an LLM (GPT-4o, GPT-5, etc.) with a chat interface, memory features, file uploads, web search, and other tools. Different companies build different LLMs; most consumer apps today use one of a handful.

Copilot Microsoft's AI product family

Microsoft's brand for AI across their products — Windows, Microsoft 365, GitHub, and more. "Microsoft 365 Copilot" is the enterprise version built into Outlook, Word, Excel, Teams, and PowerPoint.

A few flavors to know apart: Copilot (free) is consumer chat at copilot.microsoft.com. Microsoft 365 Copilot ($30/user/month) is the business version that knows your work files. GitHub Copilot is for writing code. Copilot Studio is for building custom agents. If someone says "Copilot" at a company, they usually mean Microsoft 365 Copilot.

Copilot Notebook formerly "Projects"

A Microsoft 365 Copilot feature that lets you gather specific documents into a dedicated workspace, so Copilot always answers from those files.

Think of it as a Copilot scoped to a project. Put all the files for "Pike 3400 Leasing" in one Notebook, and now every chat draws only from those documents — no noise from unrelated content. This is enterprise-grade RAG with a friendly interface. Microsoft renamed this from "Projects" to "Notebooks" in March 2026 — you may still see "Projects" in older documentation.

Fine-tuning / Customization

Taking an already-trained model and further training it on specific data to specialize it for a domain.

A medical-specialist AI might start as a general model, then be fine-tuned on medical literature. Enterprise tools like custom GPTs and AI assistants often use a lightweight version of this — or just a smart prompt — to specialize behavior without full retraining.

Open-weights / Open Source

When a model's weights are published publicly — anyone can download it and run it on their own computer.

Meta's Llama, DeepSeek, Mistral, and Alibaba's Qwen are open-weights models. You can download and run them (if you have the hardware). GPT-5, Claude, and Gemini are closed — you access them only through the provider's API. Open-weights models are driving a lot of the efficiency innovation.

Deepfake

Synthetic video, audio, or images made by AI — usually of a real person, often without consent.

Cousin-level tools can now produce convincing fakes. Used in scams, political misinformation, and ordinary content creation. Detection tools, provenance standards (like C2PA), and device-level signing are catching up — but slower than creation. If a clip makes you feel strong emotion, wait 30 seconds before sharing.

Verification "check the answer"

Checking AI's answer against a source you trust before acting on it. The single most important skill.

Verification doesn't mean checking every word. It means picking the part of the answer that will drive your decision — and confirming that part against a primary source, expert, or test. For a recipe, cook a bite. For a bill, call the biller. For a law, look at the statute. Match the rigor to the stakes.

Missing a term?

If there's an AI word you've been hearing and can't find a plain-English explanation, email me. I'll add it to this glossary and credit the ask.

Suggest a term