📖 AI Glossary
A BEGINNER FRIENDLY GUIDE TO THE MOST COMMON AI TERMS
A clear, beginner‑friendly guide to the AI terms you’ll see in ChatGPT, Claude, Gemini, Copilot, and more, with simple examples anyone can understand.
How To Use This Glossary
This glossary is here to help you feel comfortable with AI tools by offering clear, simple definitions and examples. If you’re new to AI, begin with AI Basics.
1. AI (Artificial Intelligence)
Meaning: Technology that performs tasks that normally require human thinking.
Example: AI drafts an email when you give it a short instruction.
2. LLM (Large Language Model)
Meaning: A type of AI trained on huge amounts of text so it can understand and generate language.
Example: ChatGPT, Claude, Gemini, and Copilot are all LLMs.
3. Prompt
Meaning: The message or instruction you give the AI.
Example: “Write a friendly reminder email for my client.”
4. Prompt Engineering
Meaning: Writing clear, structured prompts to get better results.
Example: Write a 150‑word LinkedIn post in a warm, confident tone about staying focused as a solopreneur.
5. Iteration
Meaning: Improving the AI’s output by giving follow‑up instructions.
Example: “Shorter.” “More formal.” “Rewrite for a beginner.”
6. Context
Meaning: Background information that helps the AI understand what you want.
Example: “This email is for a long‑time client who prefers short messages.”
7. Use Case
Meaning: A specific task you use AI for.
Example: Summarizing meetings, writing emails, and generating ideas.
8. Output
Meaning: The response the AI gives you.
Example: The email or summary the AI produces after your prompt.
9. Token
Meaning: A small chunk of text the AI uses to process language.
Example: “Hello” might be one token; longer words may be several.
10. Training Data
Meaning: The text an AI model learned from.
Example: Books, articles, websites, and other publicly available texts.
11. Models
Meaning: The specific version of an AI system you’re using.
Example: ChatGPT 4, Claude 3, Gemini Advanced, Copilot.
12. Hallucination
Meaning: When an AI confidently gives an incorrect or made‑up answer.
Example: It invents a statistic that doesn’t exist.
13. Reasoning
Meaning: The AI’s ability to analyze information and solve problems.
Example: Compare these two job candidates and explain who is a better fit.
14. System Prompt
Meaning: A behind‑the‑scenes instruction that shapes the AI’s behavior.
Example: You are a helpful assistant.
15. Tool / Integration
Meaning: A feature that lets AI interact with other apps.
Example: Uploading a PDF for the AI to summarize.
16. Model Switching
Meaning: Choosing a different AI model depending on the task.
Example: Using Claude for long text and ChatGPT for creative writing.
17. Guardrails
Meaning: Safety rules built into AI tools.
Example: The AI refuses to generate harmful or inappropriate content.
18. Temperature
Meaning: A setting that controls how creative or predictable the AI is.
Example: Low temperature → factual email
- High temperature → creative brainstorming
(Some tools hide this setting, so you may never see it.)
19. Embedding
Meaning: A way AI converts text into numbers so it can compare meaning.
Example: Used in features like “find similar documents.”
20. Fine‑Tuning
Meaning: Training an AI model on your own examples.
Example: A company fine‑tunes a model to write in their brand voice.
21. Knowledge Base
Meaning: A collection of documents that the AI can reference to give more accurate answers.
Examples:
- Uploading a PDF onboarding guide so the AI can answer questions about your process.
- Providing a folder of blog posts so the AI can match your writing style.
22. Thread / Conversation History
Meaning: The ongoing chat context that the AI uses to understand your instructions.
Example: If you say “shorter,” the AI knows you’re referring to the previous output.
23. Reset / New Chat
Meaning: Starting a fresh conversation so the AI doesn’t rely on previous context.
Example: When the AI starts drifting or misunderstanding your instructions.
24. Agent
Meaning: A more advanced AI system that can take multiple steps or use tools autonomously.
Example: An AI that can research, summarize, and draft a report without you prompting each step.
25. API (Application Programming Interface)
Meaning:
A way for different software tools to communicate with each other.
Example:
An AI tool might use an API to pull data from your calendar or documents.
26. Dataset
Meaning:
A collection of information used to train or evaluate an AI model.
Example:
A dataset of customer emails can be used to teach AI how to write replies.
27. Latency
Meaning:
The time it takes for an AI tool to respond after you submit a prompt.
Example:
If the AI takes a few seconds to load an answer, that delay is latency.
28. Bias (in AI)
Meaning:
Patterns in training data that cause the AI to produce unfair or unbalanced results.
Example:
If a model was trained mostly on formal writing, it may struggle with a casual tone.
29. Multimodal AI
Meaning:
AI that can understand or generate more than one type of input, such as text, images, or audio.
Example:
An AI that can read a PDF and also describe an image is multimodal.
30. Zero‑Shot Prompting
Meaning:
Asking AI to perform a task without giving any examples.
Example:
Write a short bio for a graphic designer” is a zero‑shot prompt.
Explore More AI Resources:
- Workflows - Step‑by‑step guides to help you use AI for real tasks
- Tool Spotlights - Clear, beginner‑friendly overviews of popular AI tools
- AI Basics - A simple introduction to how AI works and how to use it
- AI Resource Library - Browse all free guides, workflows, and tools
☑️ Return to the AI Resource Library for all guides and tools.