Classification, recommendations and unusual events
A large part of AI does not create content; it helps recognise or sort cases.
What is it?
Much of AI does not create content but sorts examples into groups, recommends products or content and detects unusual events. These systems run in the background of most online services.
An everyday example
A streaming service suggests your next series, and your bank calls because your card was suddenly used in another country.
Why does it matter?
Recommender systems shape what we read, watch and buy. Anomaly detection protects against fraud and breakdowns.
A common misconception
That a recommendation is neutral. The system chooses by a measure someone set, such as watch time or sales.
In more detail
What is classification?
The model estimates which category an example belongs to, for example “spam” or “not spam”.
What is a recommender system?
A system that ranks content or products according to estimated interest and other goals.
What is anomaly detection?
Looking for cases that differ from the usual patterns.
What is the danger with recommendations?
The system may favour clicks or holding your attention, if that is the goal it was given.
Remember: A recommendation is a choice made by a measure, not a neutral truth.
Reliable sources
Last reviewed 1 November 2026. An explainer, not professional, legal or medical advice. For rules and figures that change, check the official version.