Algorithm: a precise recipe for solving a problem
An algorithm is a finite list of precise instructions, carried out in a set order, to solve a problem. A cooking recipe or your GPS route are everyday examples, and artificial intelligence uses algorithms everywhere.
Data as of October 22, 2026
Key figures
| Level | Beginner |
|---|---|
| Introduced by | Du surnom du mathématicien Al-Khwarizmi (étymologie, Académie française) |
A series of steps, nothing more
The Académie française dictionary defines an algorithm as a method of calculation that sets out the steps to follow to solve a series of equivalent problems, by applying a finite sequence of rules in a precise order (Académie française entry, in French). In other words: clear steps, in the right order, that come to an end.
You follow algorithms every day without noticing:
- a pancake recipe: mix the flour and eggs, add the milk little by little, let the batter rest;
- sorting your mail: bills first, then personal letters, then ads;
- your GPS route, worked out by comparing possible paths to find the fastest one.
The word itself has a long history: according to the same dictionary, it goes back, via Spanish, to the Arabic Al-Khuwarizmi, the nickname of a mathematician.
Classic algorithms vs. machine learning
In a traditional program, a person writes every rule. With machine learning, you write an algorithm that learns the rules from examples instead. The outcome of that learning is an AI model. When you chat with an AI assistant, several algorithms work together: splitting your text into pieces, computing the reply word by word, filtering certain content.
Why it matters to you
Algorithms play a part in decisions that affect you: the order of job ads you see, whether a loan is approved, the videos suggested to your kids. Knowing that these are rules and calculations designed by people helps you ask the right questions: what data does it rely on? Can I challenge its decision? Does a human check it? In Europe, the GDPR gives people specific rights when decisions are made solely by automated means.
Common misconceptions
“The algorithm decides on its own.” It carries out what it was built to do. The starting choices (goal, data, criteria) are human.
“An algorithm is always neutral.” Poorly chosen data or criteria can produce bias, even with no intent to harm.
“Algorithm and artificial intelligence mean the same thing.” Every computer program relies on algorithms; artificial intelligence is one particular use of them.
To see where you stand with these ideas, take the AI level test.
Related terms
- Machine learning · A branch of AI where software learns patterns from examples instead of following rules written one by one by a person.
- 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.
- AI bias · A systematic error that makes an AI treat some groups unfairly or repeat stereotypes, often inherited from its training data.
- Artificial intelligence · Techniques that let machines perform tasks we associate with human intelligence, such as understanding, predicting or creating.
- Automation · Handing a repetitive task to software that runs it on its own, following fixed rules or with the help of AI.
Try it
Full glossary
Frequently asked questions
Where does the word algorithm come from?
According to the dictionary of the Académie française, the word goes back, via Spanish, to the Arabic Al-Khuwarizmi, the nickname of a mathematician. Its medieval French form, “augorisme”, dates from the 13th century.
Is an AI an algorithm?
AI uses algorithms, but it is not just a hand-written recipe. A learning algorithm trains a model from data; then it is the model, with its learned parameters, that produces the answers.
What is a social media “algorithm”?
It is the set of calculations that decides which posts to show you and in what order. It combines rules and machine learning models that take your past interactions into account.
Can an algorithm be unfair?
Yes, if it was poorly designed or relies on biased data. An algorithm has no intentions, but its results can disadvantage certain groups. That is why transparency and human oversight matter.
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.