Model, in AI, is the term for the system that has learned patterns from data and can make predictions or generate output. It is the trained “brain” that results from the machine learning process.
What it means in plain English
When you train a machine learning algorithm on data, the result is a model: a collection of learned patterns (stored as numbers called parameters) that maps inputs to outputs. The model is what you actually use — you feed it new input, and it produces a prediction, classification, or piece of generated content. ChatGPT, for instance, is powered by a large language model.
People often use “model” and “AI” interchangeably, but precisely, the model is the specific trained artefact that does the work.
A simple example
A spam-detection model, once trained, takes a new email as input and outputs “spam” or “not spam.” The model is the finished, usable product of the training process.
Why it matters
The model is the central object in machine learning — the thing that gets trained, evaluated, deployed, and used. Nearly every other term in this glossary relates to how models are built, what they learn from, or how they behave.
Related terms
- Parameter — the learned values that make up a model.
- Training Data — what a model learns from.
- Inference — using a trained model to make predictions.
Frequently asked questions
What is a model in AI?
A model is the output of training: a system of learned parameters that maps inputs to outputs, used to make predictions or generate content.
What is the difference between a model and an algorithm?
An algorithm is the method used to learn (like gradient descent); a model is the resulting trained artifact that you actually use to make predictions.
