Plain-English definitions of the most common AI terms. No jargon, no tech background needed.
A computer program that can learn patterns, answer questions, and complete tasks in ways that seem human-like. AI does not "think" — it predicts what a good response looks like based on billions of examples.
The process of teaching an AI by showing it millions of examples instead of programming exact rules. The more examples it sees, the better it gets.
The type of AI behind tools like ChatGPT and Claude. It was trained on huge amounts of text from the internet and books, which is why it can write, summarize, and answer questions so well.
The message or question you type to an AI tool. Writing a good prompt is the single most important skill for getting useful results from AI.
The skill of writing clear, specific prompts that get better results from AI. You do not need to be technical — just learn to be specific about what you want.
OpenAI's popular AI chatbot. You type questions or requests, it responds in natural conversation. One of the most widely used AI tools in the world.
Anthropic's AI assistant, known for being thoughtful, detailed, and safe to use. Often preferred for complex writing, analysis, and nuanced tasks.
AI that creates new content — text, images, audio, or video — based on what it has learned. ChatGPT, Claude, DALL-E, and Midjourney are all examples of generative AI.
When an AI confidently states something that is false. AI tools can make things up — always verify important facts, especially dates, names, and statistics.
The amount of text an AI can "remember" during a single conversation. When you have a very long chat, the AI may start to forget early parts of the conversation.
The massive amount of text, images, or other data that an AI was taught with. The quality of training data affects how good — and how biased — an AI's responses are.
A computer program that converses with you via text. Modern AI chatbots like ChatGPT are far more capable than old-style chatbots that followed rigid scripts.
The technology that allows computers to understand and generate human language. It is the reason you can talk to AI in plain English instead of code.
A set of rules or steps a computer follows to complete a task. When people say "the algorithm" on social media, they mean the hidden rules deciding what content you see.
A type of AI architecture loosely inspired by the human brain. It processes information through many layers of computation, which is why it's so good at recognizing patterns.
A way for one software program to talk to another. When a company builds an app powered by ChatGPT, they use the OpenAI API to connect them.
Taking a general AI and further training it on a specific type of content (like medical texts or customer service conversations) to make it better for that use case.
Using technology — including AI — to complete tasks automatically without manual effort. AI makes automation easier by understanding plain language instructions.
A type of machine learning that uses many layers of neural networks. It is what powers modern image recognition, voice assistants, and language models.
When an AI produces unfair, stereotyped, or discriminatory outputs because of patterns in its training data. A well-known problem that AI companies are actively working to reduce.
The internal "knobs" that define how an AI behaves. When you hear "a model with 7 billion parameters," more parameters generally means a more capable (and expensive) model.
A small chunk of text — roughly a word or part of a word — that AI processes one at a time. When AI providers charge by 'tokens,' they're charging based on how much text went in and out.
Zero-shot: asking AI to do something with no examples. Few-shot: giving the AI a few examples first to guide its response. Few-shot prompts often produce better results.
AI that can handle more than one type of input — text, images, audio, and video. GPT-4o and Claude 3 are examples of multimodal models.
Designing and using AI systems that are safe, fair, transparent, and beneficial to people. Includes thinking about privacy, bias, misuse, and unintended consequences.
Understanding these terms is step one. Step two is actually using AI — with a patient tutor guiding you through it.
Start Free — No Card Needed