AI conversations are full of terms that get thrown around without explanation — often assuming familiarity that not everyone has. This glossary collects the essential vocabulary from across this site into one reference, with short, clear definitions and links to full explanations for anything you want to explore more deeply.
Table of Contents
- Foundational Terms
- Machine Learning Terms
- Language Model Terms
- Generative AI Terms
- Technical Architecture Terms
- Ethics and Governance Terms
- Business and Application Terms
- How to Use This Glossary
- Frequently Asked Questions
- Key Takeaways
- Conclusion
Foundational Terms
Artificial Intelligence (AI) — The broad field of computer science focused on building systems that perform tasks normally requiring human intelligence. See our complete beginner’s guide to AI.
Narrow AI — AI designed for a specific task, such as image recognition or language translation. All AI in practical use today is narrow AI.
Artificial General Intelligence (AGI) — A theoretical, not-yet-achieved form of AI with human-level reasoning across virtually any task, discussed in our AI myths guide.
Algorithm — A set of defined steps or rules a computer follows to complete a task or solve a problem; the underlying instructions any AI system runs on.
Model — The output of a machine learning training process — a system that has learned patterns from data and can apply them to new inputs.
Machine Learning Terms
Machine Learning (ML) — A method of building AI where systems learn patterns from data rather than following explicitly programmed rules. See our machine learning guide.
Deep Learning — A subset of machine learning using multi-layered neural networks, especially effective for unstructured data like images and text. See our deep learning guide.
Neural Network — A system of connected layers loosely inspired by the brain, used to process data and learn patterns. See our neural network architecture guide.
Supervised Learning — A machine learning approach where a model learns from labeled examples (input paired with the correct answer).
Unsupervised Learning — A machine learning approach where a model finds patterns in data without predefined labels or answers.
Training Data — The dataset used to teach a machine learning model to recognize patterns.
Overfitting — When a model learns training data too specifically, including its noise and quirks, reducing its ability to generalize to new, unseen data.
Bias (AI Bias) — Systematic, unfair patterns in an AI system’s outputs, often originating from patterns in training data. See our AI bias guide.
Language Model Terms
Large Language Model (LLM) — A deep learning model trained on massive text datasets to understand and generate human language. See our LLM guide.
Token — A small unit of text (often a word or word fragment) that a language model processes. See our tokenization guide.
Context Window — The maximum amount of text (measured in tokens) a language model can consider at once in a given interaction.
Hallucination — When an AI model generates confident but factually incorrect or fabricated information.
Prompt — The input text or instructions given to an AI model to guide its response. See our prompt engineering guide.
Prompt Engineering — The practice of crafting clear, effective input to get more accurate and useful AI output.
Fine-Tuning — Further training a pretrained AI model on a smaller, specific dataset to adapt its behavior for a particular task. See our fine-tuning guide.
Embedding — A numerical representation of data (like text) positioned so that similar meanings are mathematically close together. See our embeddings guide.
Generative AI Terms
Generative AI — AI that creates new content — text, images, audio, or video — rather than just analyzing or classifying existing data. See our generative AI guide.
Diffusion Model — A generative AI architecture that creates images by gradually refining random noise into a coherent image guided by a prompt. See our AI image generation guide.
Text-to-Image — AI generation that produces an image from a text description.
Text-to-Speech (TTS) — AI technology that converts written text into spoken audio. See our AI voice and speech technology guide.
Speech-to-Text — AI technology that converts spoken audio into written text, also called speech recognition.
Deepfake — Synthetic media, often video, depicting a real person saying or doing something they didn’t actually do — a significant ethical concern covered in our AI video generation guide.
Technical Architecture Terms
Transformer — A neural network architecture that uses “attention” to weigh relationships between all parts of an input simultaneously — the architecture behind virtually all modern language models. See our transformers guide.
Attention (Self-Attention) — The mechanism transformers use to determine how relevant each part of an input is to understanding every other part.
Convolutional Neural Network (CNN) — A neural network architecture specialized for processing image and visual data. See our computer vision guide.
Natural Language Processing (NLP) — The field of AI focused on helping computers understand and generate human language. See our NLP guide.
Computer Vision — The field of AI focused on enabling machines to interpret visual information from images and video.
Retrieval-Augmented Generation (RAG) — A technique that retrieves relevant information and provides it to a language model as context before generating a response. See our RAG guide.
AI Agent — An AI system that can plan and execute multi-step tasks, using external tools, with limited step-by-step human direction. See our AI agents guide.
Multimodal AI — AI systems that can process and relate multiple data types (text, images, audio) within a single model. See our multimodal AI guide.
Vector Database — A specialized database designed to store and efficiently search embeddings based on similarity, rather than exact matches.
Ethics and Governance Terms
AI Ethics — The principles guiding fair, transparent, and accountable development and use of AI systems. See our AI ethics guide.
Responsible AI — The practical frameworks and processes organizations use to implement AI ethics principles. See our responsible AI guide.
Explainability — The degree to which humans can understand why an AI system produced a specific output or decision.
Data Privacy (in AI context) — The protection of personal information used to train, operate, and improve AI systems. See our data privacy and AI guide.
Fairness Metric — A mathematical way of measuring whether an AI system’s outcomes are equitable across different groups.
Business and Application Terms
AI Automation — Using AI to perform tasks or make decisions automatically within a workflow, adapting to variability rather than following only fixed rules. See our AI automation guide.
Chatbot — A conversational software tool, often powered by an LLM, that interprets and responds to natural language input. See our AI chatbots guide.
Predictive Analytics — Using AI to forecast future outcomes based on patterns in historical data.
Sentiment Analysis — Using natural language processing to determine the emotional tone of text.
Open Source AI Model — An AI model whose underlying code and, often, trained weights are publicly available for use and modification. See our open source vs proprietary AI guide.
How to Use This Glossary
This page is designed as a quick-reference starting point, not a replacement for the deeper guides linked throughout. If a term here sparks a question, the linked article covers the concept in full — including practical examples, common mistakes, and how it fits into the broader AI landscape.
Tip: Bookmark this page — it’s a fast way to look up a specific term without searching through longer articles, and every entry links onward if you want the fuller explanation.
Frequently Asked Questions
1. What’s the difference between AI and machine learning?
AI is the broad field; machine learning is one common method used to build AI systems, where models learn patterns from data rather than following fixed rules.
2. What does “LLM” stand for?
Large Language Model — a deep learning model trained on massive text datasets to understand and generate human language.
3. What is a token in simple terms?
A small unit of text — sometimes a whole word, sometimes a word fragment — that a language model processes as its basic unit of input and output.
4. What does “hallucination” mean in AI?
It refers to an AI model generating confident-sounding but factually incorrect or fabricated information.
5. What is the difference between generative AI and traditional AI?
Generative AI creates new content (text, images, audio); traditional AI applications more often classify, predict, or recommend based on existing data.
6. What is RAG, briefly?
Retrieval-augmented generation — a technique where relevant information is retrieved and given to a language model as context before it generates a response, improving accuracy and currency.
7. What’s the difference between fine-tuning and prompt engineering?
Fine-tuning further trains a model on specific data to adjust its behavior; prompt engineering crafts input instructions to guide an existing model’s response — no retraining involved.
8. What is a neural network, simply?
A system of connected layers, loosely inspired by the brain, that processes data through weighted connections to learn and recognize patterns.
9. What does “multimodal” mean?
It describes an AI system that can process and relate multiple types of data — like text, images, and audio — together, rather than handling only one type.
10. Where can I learn more about a specific term?
Every entry in this glossary links to a full guide covering that concept in depth, with practical examples and context.
Key Takeaways
- This glossary covers the essential vocabulary spanning foundational AI concepts, machine learning, language models, generative AI, technical architecture, ethics, and business applications.
- Each entry links to a full guide for deeper understanding.
- Understanding this vocabulary makes every other AI conversation — technical or business — significantly more accessible.
- New terms continue to emerge as the field evolves; this page reflects the terminology most relevant as of this writing.
Conclusion
AI terminology can feel like a barrier to understanding the technology itself — this glossary exists to remove that barrier, giving you quick, clear definitions alongside pathways to go deeper on anything that interests you. Bookmark it, and return whenever an unfamiliar term comes up elsewhere on this site or in the wider AI conversation.