AI Literacy / Glossary

Training data

Training data is the collection of examples, such as texts, images, sounds or past records, that an AI model learns its patterns from. What is in it, and what is missing, shapes what the model can do.

For a model that sorts photos of cats and dogs, the training data is many labelled photos. For a chatbot, it is a huge amount of text. The model can only learn patterns that appear in its examples.

If some groups, situations or languages are rare in the training data, the model tends to work worse for them. This is one of the main ways bias gets into AI.

Training data can also include personal information or copyrighted work, which is why there are legal debates and rules about how it is collected and used. Many chat services say in their settings whether your chats may be used to train future models.