In one sentence: A large language model, or LLM, is the kind of AI behind many chatbots, trained on a very large amount of text to predict which words fit.
Why it matters: LLMs can write convincing text, but they can be wrong. Companies may store what you type, so it helps to read their terms of use (the agreement you accept by using the service) and privacy policy.
One thing to do: Find the privacy page of one AI service you use and look for how it uses your chats.
Related terms: chatbot, training data, hallucination
Checked: 25 Oct 2026.
