Essential AI Terms for Beginners

The 14 terms

Prompt

The instruction or question you type to an AI tool. A more specific prompt — with purpose, audience, and format spelled out — reliably gets a more useful response than a vague one.

LLM (large language model)

The underlying technology behind tools like ChatGPT and Claude — a model trained on huge amounts of text that generates responses by predicting what text should come next, rather than looking answers up in a database.

Hallucination

When an AI states something false with the same confident tone as something true — a made-up citation, a wrong date, a fabricated fact. It's a real limitation worth knowing about, not a rare glitch.

Token

A chunk of text — often close to a word or part of a word — that an AI model processes as its basic unit. Usage limits and context limits are usually measured in tokens rather than words or characters.

Context window

The amount of text (measured in tokens) an AI model can consider at once in a conversation, including everything you've said so far. Once a conversation exceeds it, earlier parts start getting dropped from what the model can see.

System prompt

A standing instruction set once at the start of a conversation or set up in an account's custom instructions that shapes every response — for example, telling the AI your role, your preferred tone, or a rule to always follow.

Temperature

A setting (mostly relevant in developer tools, not typical chat interfaces) that controls how predictable versus varied an AI's output is. Lower temperature gives more consistent, conservative answers; higher gives more varied, creative ones.

Fine-tuning

The process of further training an existing AI model on a specific, narrower set of data so it performs better on a particular task or in a particular style — distinct from just prompting the base model well.

Multimodal

An AI model that can work with more than just text — for example, reading an image, listening to audio, or generating an image — rather than being limited to text in, text out.

Agent

An AI system set up to take a series of actions toward a goal with limited step-by-step supervision, rather than just responding once to a single prompt — for example, researching a topic across multiple steps and compiling the findings on its own.

RAG (retrieval-augmented generation)

A technique where an AI model looks up relevant information from a specific set of documents or a database before generating its answer, so the response is grounded in that specific source material instead of only the model's training data.

Few-shot

A prompting technique where you include a few examples of the input/output pattern you want directly in the prompt, so the AI can match that pattern rather than guessing at your intent from a description alone.

Fine print: model

The specific underlying AI system you're talking to (for example, a particular version of GPT or Claude) — different models from the same company can vary meaningfully in ability, speed, and cost.

Custom instructions

A settings feature (on ChatGPT and similar tools) that lets you save standing context about yourself and your preferences once, so you don't have to repeat it in every new conversation.

Which ones actually matter day-to-day

The AI Beginner Courses channel on YouTube uses several of these terms naturally throughout its videos, which is a reasonable way to absorb them in context. For everyday use, two terms do the most practical work: hallucination (know that specific facts need a quick check) and context window (know that very long conversations eventually lose track of earlier detail, so summarizing or starting fresh sometimes helps). Prompt, system prompt, and custom instructions matter next, since they directly shape the quality of what you get back. The rest — token, temperature, fine-tuning, multimodal, agent, RAG, few-shot — are more useful for understanding AI news and product descriptions than for operating a chat interface day to day.

Where to go deeper

This list is intentionally short — 14 terms, not 40. For a complete reference covering the rest of the AI vocabulary you'll run into (including newer, more specialized terms), see the full AI Glossary. And for how these terms actually apply once you're writing real prompts, see how to write your first AI prompt.