Large Language Model

AI & Generative Search

Also: LLM · Large Language Models

What it isAI trained to predict and generate text

PowersChatGPT, Gemini, Claude, AI Overviews

Marketing useContent, search answers, chat, analysis

Quick definition

A large language model, or LLM, is an artificial intelligence system trained on enormous volumes of text to predict the next word in a sequence. That simple mechanism, run at massive scale, lets it write, summarise, answer questions and hold a conversation. LLMs are what sit behind ChatGPT, Google Gemini, Claude and the AI Overviews now appearing in search results.

Where it shows up in the data

Next-token prediction

The core mechanism. The model predicts the most likely next piece of text given everything before it, one token at a time. Scale makes this feel like reasoning.

Training data

The text the model learned from. It shapes what the model knows, how it writes and the biases it carries. Most models have a knowledge cutoff date beyond which they know nothing.

Hallucination

When a model generates plausible-sounding text that is false. It is not lying, it is predicting likely words with no fact check, which is why human review matters.

What it actually means

An LLM is trained by reading a very large corpus of text and learning the statistical patterns of language. Given a prompt, it generates a response one token at a time, each choice shaped by everything it read in training plus the conversation so far. It has no database of facts and no understanding in the human sense. It produces text that is statistically likely to be a good continuation. That is why it can write a fluent marketing brief and also invent a citation that does not exist. For marketers the important shift is that LLMs increasingly stand between your content and your audience, summarising, recommending and answering on your behalf.

A large language model does not know facts. It predicts likely words. That is why it sounds confident and is sometimes confidently wrong.

Where people get this wrong

Trusting factual output without checkingAn LLM predicts likely text, it does not verify facts. It will invent statistics, sources and quotes that look real. Every factual claim needs checking before it goes out.

Publishing raw model output as contentUnedited LLM text is generic, often inaccurate and increasingly penalised by search engines rewarding genuine expertise. Use it to draft, then add real insight and voice.

Ignoring LLMs as a distribution channelPeople now ask ChatGPT and Gemini the questions they used to type into Google. If your business is not part of the answer, you are invisible in a growing slice of search.

Common questions

What is a large language model in plain English?

It is an AI trained on huge amounts of text to predict what word comes next. Done at scale, that lets it write, answer questions and hold a conversation. ChatGPT and Gemini are built on large language models.

Can I trust what a large language model tells me?

Not without checking. LLMs predict likely-sounding text rather than retrieving verified facts, so they can state false things confidently. Use them for drafts and ideas, then verify anything factual.

Why do large language models matter for marketing?

They speed up content and analysis work, and they are becoming a place people search. When an LLM answers your customer's question, you want your business to be the source it draws on.

Debrief

Get the next one

No spam. No fluff. Just the next article, straight to your inbox.

Keep exploring

About New Rebellion

New Rebellion is a marketing intelligence consultancy. We build tools, score Australian businesses on how their marketing actually performs, and publish Debrief every day. This dictionary is part of how we work in the open.

How we think →