AI Glossary — Key AI Terms & Definitions

A

Agentic AI

AI systems that can autonomously plan, take multi-step actions, use tools, and achieve goals without step-by-step human instruction.

A

AI Agent

A software system powered by an AI model that autonomously perceives its environment, makes decisions, and takes actions to achieve a specified goal.

A

AI Alignment

The field of research and engineering focused on ensuring AI systems pursue goals and behave in ways that match human values, intentions, and safety expectations.

A

AI Benchmark

A standardized test or dataset used to measure and compare the capabilities of AI models on specific tasks like knowledge, coding, math, or reasoning.

A

AI Hallucination

When an AI model generates information that sounds confident and plausible but is factually incorrect, fabricated, or not supported by its sources.

A

AI Reasoning Model

A class of language models trained to 'think' before answering — generating extended internal reasoning steps to solve complex problems more reliably.

C

Context Window

The maximum number of tokens (words and punctuation) a language model can process at once in a single prompt plus response.

E

Embeddings

Dense numerical vector representations of text, images, or other data that capture semantic meaning in a form machines can compare.

F

Fine-Tuning

Training a pre-trained language model on a smaller, task-specific dataset to improve its performance on a particular domain or task.

F

Function Calling

A capability that lets a language model request execution of predefined functions or tools by outputting structured arguments, enabling AI to take real actions.

I

Inference

The process of running a trained AI model on new inputs to produce predictions or generated outputs, as opposed to the training phase.

K

Knowledge Graph

A structured representation of entities (people, places, concepts) and the relationships between them, stored as a graph for reasoning and retrieval.

L

Latent Space

The compressed, high-dimensional mathematical space in which an AI model represents the essential features and relationships of its data.

L

LLM (Large Language Model)

A deep learning model trained on vast quantities of text that can generate, summarise, translate, and reason about language at human level or above.

M

MCP (Model Context Protocol)

An open standard by Anthropic that lets AI assistants connect to external tools, data sources, and services through a unified protocol.

M

Mixture of Experts (MoE)

A model architecture that contains many specialized sub-networks (experts) but activates only a few per input, delivering large-model capability at lower compute cost.

M

Multimodal AI

AI systems that can process and generate content across multiple data types — text, images, audio, and video — within a single model.

O

Open Weights

AI models whose trained parameters (weights) are publicly released, allowing anyone to download, run, modify, and fine-tune the model locally.

P

Prompt Engineering

The discipline of designing and refining input prompts to elicit more accurate, reliable, and useful responses from AI language models.

R

RAG (Retrieval-Augmented Generation)

A technique that combines a language model with real-time document retrieval to produce grounded, up-to-date answers.

S

Small Language Model (SLM)

A compact language model — typically under ~15 billion parameters — designed to run efficiently on local or edge hardware while remaining capable for focused tasks.

S

Synthetic Data

Artificially generated data that mimics the statistical properties of real data, used to train AI models when real data is scarce, private, or imbalanced.

T

Tokenization

The process of breaking text into smaller units called tokens — sub-words, characters, or symbols — that a language model can process numerically.

T

Transformer Model

The neural network architecture, based on self-attention, that underpins virtually all modern large language models and many vision and audio models.

V

Vector Database

A database purpose-built to store and query high-dimensional vector embeddings, enabling semantic similarity search at scale.

The AI Glossary — Plain English

Essential AI Terms