Lesson 1 of 6
What do we mean by AI?
A message goes into a chat application and a reply comes back. This lesson names the parts that made that possible, so the rest of the module can explain how each one works.
Numbered links identify sources. Underlined terms offer a definition without leaving the lesson.
A useful starting point
Artificial intelligence, usually shortened to AI, is a broad area of computing concerned with abilities such as recognizing patterns, interpreting language and making decisions. The term does not name one company, one product or one method. Recognizing speech, estimating tomorrow’s rainfall and generating a picture are different tasks, so a claim about one kind of AI does not automatically apply to another. This course concentrates on systems that work with language, because they are a foundation of many systems that carry out work, which later modules explain.
Three words describe the parts of this interaction. Software is a set of instructions that a computer carries out. An , or app, is software arranged so that a person can use it: a messaging app, a calendar, a chat window. Data is the information the software works with: a sentence you type, a photograph, a list of opening times.
Software can follow rules that people write explicitly. A reminder app shows a message when the clock reaches the time you chose, because someone wrote an instruction saying exactly that. With machine learning, people take a different route. Instead of writing every rule, they collect examples and use them to develop a model that can recognize patterns or produce predictions. The computer still performs ordinary calculations. What changes is that some of the numbers steering those calculations were adjusted using data rather than typed in by a person. That is the kind of AI this module explains. 1
The model is one part of the application
A is the part that uses what it learned during training to turn an input into an output. For now, “input” means the information given to it and “output” means what it produces. The next lesson explains how the learning happens.
The application and the model do different jobs. The application draws the chat screen, accepts your message, passes information to the model and shows what comes back. The model produces the wording of the reply. Because these are separate pieces of software, either can change while the other stays the same. A company can redesign the screen, add a dark theme or move the send button without touching the model. It can also swap in a newer model while the screen looks exactly as it did yesterday. So a visible change in the app is not, by itself, evidence that the model changed. To know whether the model changed, you need information about the model, such as the version the product says it is using.
There are also organizations doing this work. An AI lab is an organization or team that researches and develops AI, including training models. OpenAI, for example, develops models and offers ChatGPT, an application that uses them. The company, its models and its application are related, but they are three different things, and this module keeps them apart. 6
Three different roles
- 01OrganizationDevelops or supplies the system
- 02ModelProduces output from supplied input
- 03ApplicationPresents the interface and other capabilities
One example we will return to
Imagine you help organize a community garden. A neighbor asks when the next volunteer session starts. You have this note:
Saturday: meet at 10 a.m. by the north gate. Bring gloves. The session is cancelled if it rains.
You paste the note into a chat application and ask for a friendly reply. It might produce:
Come along on Saturday at 10 a.m. We’ll meet by the north gate. Bring gloves, and remember that we’ll cancel if it rains.
This reply is an authored illustration, not the output of a tested model. Every garden message in this module is a teaching example written for the lesson.
Name the parts. The note is data. The application accepted your request and showed the result. The model produced the wording. Notice what the example does not show: nothing here sent the message to the neighbor, and nothing checked the weather. Those would be further actions, and lesson six explains what software would need in order to take them.
One thing still needs explaining. Nobody typed this exact reply into the application in advance, yet the application produced it. How can a computer produce wording no one wrote? The answer is what “learning” means in machine learning.
Carry this forward
- AI names a broad field; this course starts with language-based systems built by machine learning.
- Data is the information software works on. A model uses what it learned during training to produce an output from an input.
- The organization, the model and the application are separate parts. A visible change to the app is not evidence of a change to the model.
Lesson check
Choose an answer for each question.
This is a practice check. Answers reset when you leave this page. Results are not saved and do not place you in another module or award a qualification.
The parts have names. Next, we explain how training produces a model that can respond to information it has never seen.
Sources · Reviewed 2026-09-11
Sources support nearby explanations. Product details can change. Scenarios and model results are authored illustrations unless a lesson explicitly identifies a recorded execution. Module 7 identifies its offline code checks.