The AI glossary, in plain English
No jargon, no hype. Clear definitions of the AI terms that actually come up, each with a link to learn it hands-on.
Large Language Model (LLM)
A large language model is an AI system trained on huge amounts of text to predict the next word, which lets it write, summarize, and answer in natural language.
Prompt Engineering
Prompt engineering is the skill of writing instructions that get an AI to produce the output you actually want, by giving it context, a clear task, and constraints.
Tokens (in AI)
A token is a chunk of text (roughly a word or part of a word) that an AI reads and writes in. Models measure input, output, and cost in tokens.
AI Hallucination
An AI hallucination is when a model states something false as if it were true, like a fake citation or invented fact, because it predicts plausible text rather than verified facts.
RAG (Retrieval-Augmented Generation)
RAG, or retrieval-augmented generation, is a technique where an AI first retrieves relevant documents and then answers using them, so its response is grounded in real sources instead of memory alone.
AI Agent
An AI agent is an AI system that can take actions toward a goal, not just answer, by planning steps and using tools like search, code, or other software.
Context Window
A context window is the maximum amount of text (measured in tokens) an AI can consider at one time, including your prompt, any documents, and its own answer.
Fine-Tuning
Fine-tuning is further training a model on your own examples so it learns a specific style or task, changing how the model behaves rather than just what you tell it in a prompt.
Few-Shot Prompting
Few-shot prompting means giving the AI a few examples of what you want before your actual request, so it copies the pattern instead of guessing from a description.
System Prompt
A system prompt is a behind-the-scenes instruction that sets an AI's role, rules, and tone for a whole conversation, shaping how it responds before you type anything.
Temperature (AI)
Temperature is a setting that controls how predictable or varied an AI model's responses are — low values produce focused, consistent output; high values produce more creative, unpredictable output.
Embeddings
Embeddings are numerical representations of text (or images, audio, etc.) that capture meaning — so that similar concepts end up close together in mathematical space, even if the words are different.
Vector Database
A vector database stores numerical representations of content (embeddings) and lets you search them by meaning — finding the closest match to a query, even when the exact words don't overlap.
Multimodal AI
Multimodal AI refers to models that can work with more than one type of input or output — for example, a model that accepts both text and images, or one that can read a chart and answer questions about it.
Chain-of-Thought Prompting
Chain-of-thought prompting is a technique where you ask an AI to reason through a problem step by step before giving its final answer, which tends to produce more accurate results on complex or multi-step tasks.
Zero-Shot Prompting
Zero-shot prompting means giving an AI a task or question with no examples — you rely entirely on the model's existing knowledge and your instructions to get the result you want.
Transformer
A transformer is the type of neural network architecture that underlies virtually all modern AI language models — it's the design that made models like GPT, Claude, and Gemini possible.
Parameters (AI Model)
Parameters are the internal numerical values that an AI model learns during training — they encode everything the model knows, and the number of parameters is the standard way to describe a model's size.
Training Data
Training data is the large collection of text, images, or other content that an AI model was exposed to during training — it's the raw material the model learned its knowledge and capabilities from.
Knowledge Cutoff
A knowledge cutoff is the date up to which an AI model's training data was collected — the model has no direct knowledge of events, publications, or changes that occurred after that date.
Prompt Injection
Prompt injection is a type of attack where hidden or unexpected instructions in content fed to an AI cause it to override its original instructions and do something unintended.
Reasoning Model
A reasoning model is an AI model specifically trained to think through problems in extended internal steps before producing a final answer, making it significantly more accurate on complex analytical, math, or multi-step tasks.
Function Calling / Tool Use
Function calling (also called tool use) is a capability that allows an AI model to trigger predefined actions — like running a database query, calling an API, or searching the web — rather than only producing text.
AI Guardrails
AI guardrails are rules, filters, and constraints applied to an AI system to prevent it from producing harmful, off-topic, or otherwise unacceptable outputs for its intended use case.
Inference
Inference is the process of running a trained AI model on a new input to produce an output — it's what's happening computationally every time you send a prompt and receive a response.
Learn it, or have it done for you
Definitions are the start; the course is how you use any of it. Start the course free and build a working AI habit yourself — or, if you'd rather skip to the outcome, MCF Agentic builds the AI workflows into your business directly.