Training data
Training data is the collection of examples a model learns from: text, images, recordings, numbers. Its size, quality, language mix and sources largely decide what the model is good at and what it gets wrong.
In one line, for a 12-year-old
Training data is the giant pile of examples an AI studies before it can do anything.
An example
A model trained mostly on English web pages will usually write better English than French, and may know little about life in small Canadian towns.
Why it matters to people
Training data raises hard questions: was it collected with permission, does it include personal information, and whose voices are missing? Courts and regulators in several countries are now weighing in.