Key Takeaways
- Most modern AI is built with machine learning, which finds patterns in example data instead of following hand-written rules.
- Large language models generate text by predicting likely sequences of words; tools like ChatGPT, Claude, and Gemini are built on them.
- Prompts, tokens, and the context window shape what an AI tool can do in a single conversation.
- AI systems can 'hallucinate' false facts, so important information should be verified against reliable sources.
Artificial intelligence now shows up in phones, search engines, email, and customer-service chats. The vocabulary around it can make simple ideas sound complicated. This glossary explains the terms that come up most often, in plain English, so readers can follow the conversation without a technical background.
What is artificial intelligence (AI)?
Artificial intelligence is a broad label for computer systems that perform tasks people usually associate with human thinking, such as recognizing speech, sorting images, translating text, or answering questions. Most AI in everyday use is narrow: it is built for a specific kind of task rather than general reasoning about the world.
What is machine learning?
Machine learning is the most common way modern AI is built. Instead of programmers writing a rule for every situation, the system is shown large amounts of example data and adjusts itself to find patterns. A spam filter, for example, learns what spam tends to look like from many labeled examples rather than from a hand-written list of every spam message.
What is a large language model (LLM)?
A large language model is a machine-learning system trained on a very large collection of text. It learns statistical patterns in how words and ideas tend to follow one another, which lets it generate new text, summarize documents, answer questions, and draft writing. Tools such as ChatGPT, Claude, and Gemini are built on large language models.
What is generative AI?
Generative AI refers to systems that create new content, such as text, images, audio, or code, rather than only sorting or labeling existing content. A large language model writing a paragraph and an image tool producing a picture from a description are both examples of generative AI.
What is a chatbot?
A chatbot is a program that holds a conversation through text or voice. Older chatbots followed fixed scripts. Many newer chatbots are powered by large language models, which allows them to respond to a much wider range of questions in natural language.
What is a prompt?
A prompt is the instruction or question a person gives an AI tool. Clear prompts usually produce more useful results. Stating the goal, the audience, and the desired format (for example, a short summary or a bulleted list) tends to help.
What are tokens?
Language models read and write text in small pieces called tokens. A token is often a word or part of a word. Tokens matter because AI services usually measure usage and set limits based on how many tokens go in and come out.
What is a context window?
The context window is the amount of text a language model can consider at one time, including the conversation so far and any documents provided. When a conversation or document is longer than the context window, the model cannot take all of it into account at once.
What is training data?
Training data is the collection of examples used to build a machine-learning model. The quality and coverage of that data shape what the model does well and where it falls short. A model may also have a training cutoff, meaning it was not trained on information published after a certain date unless it is connected to a live source.
What is an AI hallucination?
A hallucination is when an AI system states something false with confidence, such as an invented fact, quote, or citation. It happens because language models generate likely-sounding text rather than looking up verified answers. Hallucinations are one of the most important reasons to check AI output against reliable sources.
Using AI tools sensibly
- Verify before relying on it. Important facts, figures, and citations from an AI tool should be checked against an original, reliable source.
- Be careful with sensitive information. Many AI services may store or review what users type. Medical details, legal documents, account numbers, and other private information are best kept out of public AI tools.
- Know the limits. AI tools can explain general concepts, but they do not know the facts of a specific situation and are not a substitute for advice from a qualified professional.
For a deeper look at how these systems are built, see How Does AI Actually Work? A Beginner's Guide.
This article is general educational information about technology. It is not legal advice.