Why you need this
AI conversations are full of jargon that can make you feel left behind. Colleagues mention "fine-tuning" and "RAG," news articles reference "hallucinations" and "transformers," and job postings ask for "prompt engineering" skills—but nobody stops to explain what these terms actually mean.
The problem: AI vocabulary evolves fast. New terms appear monthly, and even people who work with AI daily struggle to keep up. Nodding along when you don't understand wastes your time and can lead to poor decisions about AI tools and strategies.
This vocabulary builder solves that. It covers 100 essential AI terms with plain-English definitions, real-world examples, and pronunciation guides so you can confidently join any AI conversation—whether that's a team meeting, a tech conference, or a casual chat with friends.
Perfect for:
- Anyone hearing AI terms they don't understand
- Professionals evaluating AI tools for their team
- Students learning about AI for the first time
- Managers who need to communicate with technical teams
- Career changers moving into AI-adjacent roles
What's inside
100 Terms Across 6 Categories
Core Concepts (15 terms):
- Artificial Intelligence (AI) — Machines that can perform tasks normally requiring human intelligence
- Machine Learning (ML) — AI that learns from data rather than following explicit rules
- Neural Network — Computing system inspired by the human brain's structure
- Deep Learning — ML using neural networks with many layers for complex pattern recognition
- Natural Language Processing (NLP) — AI that understands and generates human language
- And 10 more foundational terms...
Language Models (20 terms):
- Large Language Model (LLM) — AI trained on massive text data to understand and generate language
- Prompt — The instruction you give to an AI to get a response
- Token — The basic unit an LLM processes (roughly 3/4 of a word in English)
- Context Window — How much text an AI can consider at once (measured in tokens)
- Temperature — A setting that controls how creative or random AI responses are
- Hallucination — When AI confidently generates false or made-up information
- And 14 more language model terms...
Training & Learning (18 terms):
- Training Data — The information used to teach an AI model
- Fine-Tuning — Customizing a pre-trained model for a specific task or domain
- RLHF (Reinforcement Learning from Human Feedback) — Training AI using human preferences
- Supervised Learning — Teaching AI with labeled examples (input + correct answer)
- And 14 more training terms...
Advanced Techniques (20 terms):
- RAG (Retrieval-Augmented Generation) — Giving AI access to external knowledge to improve accuracy
- Embeddings — Converting text into numbers that capture meaning for comparison
- Vector Database — A database optimized for searching by meaning rather than keywords
- Agent — An AI system that can take actions and use tools autonomously
- And 16 more advanced terms...
Performance & Evaluation (15 terms):
- Benchmark — A standardized test to compare AI model performance
- Bias — Systematic unfairness in AI outputs, often reflecting training data prejudices
- Accuracy — How often an AI's responses are factually correct
- Guardrails — Safety filters that prevent AI from generating harmful content
- And 11 more evaluation terms...
Practical Terms (12 terms):
- API — A way for programs to communicate with AI services
- Inference — When an AI model generates a response (the "thinking" part)
- Deployment — Putting an AI model into production for real users
- Latency — The time delay between sending a prompt and receiving a response
- And 8 more practical terms...
Each Term Includes:
Plain-English Definition: No jargon allowed—every definition uses everyday language
Pronunciation Guide: For terms like "RLHF" (arr-ell-aitch-eff) and "Llama" (lah-muh)
Real-World Example: How you'd encounter this term in practice
Related Terms: What to learn next to build connected understanding
Why It Matters: Practical significance—why you should care about this concept
Sample terms
Term: Hallucination
- Definition: When AI confidently states something that's completely false or made up. The AI isn't lying—it's generating plausible-sounding text that happens to be wrong.
- Example: You ask AI "Who wrote the novel 'The Azure Coast'?" and it confidently names an author and publication date—but the book doesn't exist. The AI constructed a realistic-sounding answer from patterns in its training data.
- Why it matters: Understanding hallucinations helps you know when to verify AI outputs (always for facts, dates, and quotes) and when AI is generally reliable (creative writing, brainstorming, code structure).
Term: RAG (Retrieval-Augmented Generation)
- Pronunciation: "rag" (like the cloth)
- Definition: A technique where AI searches through your documents or data to find relevant information before answering, instead of relying only on what it learned during training.
- Example: A company's AI chatbot uses RAG to search their help docs when you ask a question, so it gives accurate answers about their specific products rather than generic information.
- Why it matters: RAG is how businesses give AI access to their private data without retraining the entire model. It's the technology behind most "chat with your documents" tools.
How to use it
- Browsing: Read through one category at a time to build foundational knowledge
- Reference: Look up unfamiliar terms when you encounter them in articles, meetings, or conversations
- Quiz yourself: Use the self-test section to check which terms you've mastered
- Team training: Share with colleagues to build shared AI vocabulary across your organization
- Interview prep: Review before AI-related job interviews or presentations
Want to go deeper?
This vocabulary builder gives you definitions. For deeper understanding of key concepts:
- Guide: What is AI? — The complete beginner's introduction
- Guide: Understanding ChatGPT — How the most popular AI tool works
- Guide: NLP Basics — Understanding language AI
- Full Glossary — Detailed explanations of every term with examples
License & Attribution
This resource is licensed under Creative Commons Attribution 4.0 (CC-BY). You're free to:
- Use for personal learning and reference
- Share with teams, students, and colleagues
- Print for workshops or training sessions
Just include this attribution:
"AI Vocabulary Builder" by Field Guide to AI (fieldguidetoai.com) is licensed under CC BY 4.0
Access now
Ready to build your AI vocabulary? View the complete resource online—no signup or email required.