Machine learning: learning by example
Machine learning is the way a computer learns from examples instead of following rules a person wrote by hand. It powers almost every AI you use, from your spam filter to ChatGPT.
Data as of October 22, 2026
Key figures
| Level | Beginner |
|---|---|
| Origin | 1959 |
| Introduced by | Arthur L. Samuel (IBM) |
Learning instead of following rules
A traditional program follows instructions written by a person: “if the email contains this word, mark it as spam.” The trouble is that the real world is too messy to anticipate everything. Machine learning flips the logic: you show the program a large number of examples that are already sorted (emails labeled “spam” or “not spam”) and it works out for itself which clues separate the two.
Those examples are the training data. The result of the learning process is an AI model, which then makes decisions on cases it has never seen.
The idea is not new. In 1959, IBM researcher Arthur Samuel published a paper about a checkers program that learned to play better than its own author (Arthur Samuel’s 1959 paper). That is where the phrase “machine learning” took hold.
Where you already meet it
- Your inbox filters out junk mail.
- Your bank flags an unusual card payment and sends you an alert.
- Your phone camera detects faces and adjusts the focus.
- Streaming apps suggest a song or show close to your taste.
- An AI assistant like ChatGPT runs on a model trained with machine learning at a massive scale.
Why it matters to you
Understanding this principle changes how you read an AI’s answers. A system that learns from examples reproduces what it has seen: if the examples are incomplete or lopsided, so are its decisions. That is where AI bias comes from. It also explains why an AI can be confidently wrong: it generalizes, it does not “know” things the way you do.
Common misconceptions
“The machine learns on its own, with no humans involved.” People choose the data, the method and the goal, and they check the results. The learning is automatic; the design is not.
“Machine learning and deep learning are the same thing.” Deep learning is one family of machine learning techniques, built on neural networks with many layers. Not all of AI uses it, but today’s assistants do.
To see these milestones in context, browse the AI timeline.
Related terms
- Artificial intelligence · Techniques that let machines perform tasks we associate with human intelligence, such as understanding, predicting or creating.
- Deep learning · Machine learning that uses neural networks with many layers; it is the technique behind today's chatbots, image and voice AI.
- Neural network · Software made of many small connected computing units that learns by adjusting the strength of the links between them.
- Training data · The examples (text, images, audio and more) an AI model learns from; they shape what the model knows and where it goes wrong.
- Algorithm · A finite list of precise steps, followed in a set order, to solve a problem or complete a task, like a recipe for a computer.
- AI model · The trained “brain” behind an AI app: a large file of learned numbers that turns an input into an output, such as a reply or an image.
Try it
Full glossary
Frequently asked questions
What is the difference between AI and machine learning?
Artificial intelligence is the broad field: getting machines to do tasks that seem to need intelligence. Machine learning is one way to get there, by learning from data. Today it is by far the dominant approach.
Who coined the term machine learning?
It is credited to Arthur Samuel, an IBM researcher, who published a 1959 paper on a checkers program that improved by playing. The paper appeared in the IBM Journal of Research and Development.
Does an AI keep learning while I use it?
Usually not in real time. A model is trained, then frozen before release. Your conversations may, however, be used to train a future version unless you opt out in the settings.
Do I need to code to use machine learning?
Not to use it: you already do through your phone, your inbox or an AI assistant. To build your own models, some coding helps, although no-code tools exist.
Data sources
Data as of October 22, 2026
Report an error
Does a figure, a date or a fact look wrong? Tell us: every report is checked.