AI has a jargon problem, and it gets in the way of actually using these tools. This glossary cuts through it. Below are 30 terms you will run into constantly in 2026, each defined in plain English with no hand-waving. Grouped by theme so related ideas sit together. Bookmark it and come back whenever a word trips you up.
The basics
Artificial Intelligence (AI) — Software that performs tasks we normally associate with human thinking, like understanding language or recognizing images.
Machine Learning (ML) — A way of building AI by letting a program learn patterns from data instead of being given explicit rules.
Model — The trained program that does the actual work. When people say "ChatGPT's model," they mean the underlying system that generates responses.
Training — The process of feeding a model huge amounts of data so it learns patterns. This happens once, in advance, and is expensive.
Algorithm — A set of step-by-step instructions a computer follows to solve a problem.
Neural network — A model design loosely inspired by the brain, made of layers of simple units that pass signals along to recognize patterns.
Language models
Large Language Model (LLM) — A model trained on massive amounts of text to predict and generate human-like language. ChatGPT, Claude, and Gemini are LLMs.
Token — A chunk of text the model processes, roughly a word or part of a word. Usage and limits are often counted in tokens.
Context window — How much text a model can "hold in mind" at once. A bigger window means it can read longer documents without forgetting the start.
Parameters — The internal numbers a model adjusts during training. More parameters can mean more capability, but not always.
Knowledge cutoff — The date after which a model has no built-in knowledge, because its training data stops there.
Generative AI — AI that creates new content — text, images, audio, code — rather than just classifying or analyzing.
Using AI day to day
Prompt — The instruction or question you give an AI. Better prompts get better answers.
Prompt engineering — The skill of wording prompts to get reliable, high-quality results.
System prompt — A hidden instruction that sets a chatbot's behavior and personality before you ever type anything.
Temperature — A setting that controls randomness. Low temperature gives focused, predictable answers; high gives more varied, creative ones.
Fine-tuning — Taking an existing model and training it a bit more on specialized data so it performs better for a specific task.
Inference — The act of the model actually generating a response when you use it (as opposed to training it).
Capabilities and add-ons
Multimodal — A model that handles more than one type of input or output, such as text plus images or audio.
RAG (Retrieval-Augmented Generation) — A technique where the AI looks up relevant documents first, then writes an answer based on them — reducing made-up facts.
Agent — An AI that can take actions on its own, like using tools, searching the web, or completing multi-step tasks, not just chatting.
API — A way for developers to plug an AI model into their own apps and software.
Embedding — A way of turning text into numbers that capture meaning, so software can find related content. It powers most AI search.
Chain of thought — When a model reasons step by step before answering, which usually improves accuracy on hard problems.
Risks and limits
Hallucination — When an AI confidently states something false. The single most important limitation to remember.
Bias — Unfair patterns a model picks up from its training data, which can show up in its answers.
Guardrails — The safety rules built in to stop a model from producing harmful or disallowed content.
Deepfake — Realistic fake images, audio, or video generated by AI, often used to impersonate real people.
Overfitting — When a model learns its training data too literally and performs poorly on anything new.
AGI (Artificial General Intelligence) — A hypothetical AI that matches human ability across virtually any task. It does not exist yet, despite the hype.
Takeaway
You do not need to memorize all 30 of these to use AI well. The ones that matter most in everyday use are prompt, context window, knowledge cutoff, and especially hallucination — knowing those four will keep you out of most trouble. Keep this page handy and the rest will sink in naturally as you go.
Practical, no fluff. Subscribe for one AI tip every week.
Comments
Post a Comment