You have probably typed a question into ChatGPT and gotten a surprisingly good answer back. But what is actually happening? Is it searching the web? Looking things up in a database? Thinking?
Not quite. Let me explain what ChatGPT actually is.
Think of it like a very well-read autocomplete
You know how your phone predicts the next word when you are texting? ChatGPT works on a similar idea, but taken much further.
Instead of predicting one word at a time from a small vocabulary, the model behind ChatGPT has learned from an enormous amount of text: books, articles, websites, code, conversations. It learned patterns from all of that reading. When you ask ChatGPT something, that model uses those patterns to generate a response that fits your question.
It is not looking anything up. It is not thinking in the way a person thinks. It is producing text based on what it learned tends to follow what you wrote.
So what is it exactly?
ChatGPT is a chatbot built on top of a type of AI called a large language model, or LLM. The model is the engine that learned from all that text. The chatbot is the interface that lets you have a conversation with it.
That distinction matters. ChatGPT is OpenAI’s chatbot. Other companies make similar tools too, like Claude from Anthropic and Gemini from Google. They are not all ChatGPT, but they are all part of the same bigger category: AI chatbots powered by large language models.
The name breaks down like this:
| Part | What it means |
|---|---|
| Chat | It is designed as a conversation |
| GPT | Generative Pre-trained Transformer (the model type underneath) |
Do not worry about what GPT stands for yet. You will get to that in a later lesson. The important thing now is: ChatGPT is a program that generates text based on patterns it learned.
What it is good at
ChatGPT is excellent at:
- Explaining ideas in plain language
- Drafting emails, essays, and messages
- Summarising long documents
- Writing and debugging code
- Brainstorming and iterating on ideas
What it is not
ChatGPT is not a search engine. It does not look things up in real time (unless it has a browsing tool attached). It is not always correct. It can sound confident while being wrong. You will learn more about that in a later lesson on hallucination.
It is also not conscious. It does not have opinions, feelings, or goals. It produces text that sounds like it does, because that is what the patterns it learned look like.
ChatGPT is OpenAI’s chatbot built on a large language model. It is one example in a wider group of AI chatbots, alongside tools like Claude and Gemini. It generates text by predicting what comes next based on patterns learned from a huge amount of text. It is not searching the web or thinking like a person.