What is training data?
Training data is the collection of examples an AI model learns from, such as labelled photos, text or past decisions. The model can only be as good, and as fair, as its training data: gaps or bias in the data become gaps and bias in the model.
How does training data work?
During training, the model sees the examples many times and adjusts itself to reproduce the patterns in them. Data is usually split so some is held back to test the model fairly.
Why does training data matter?
Most AI failures trace back to data. Asking 'what was this trained on?' is the quickest way to judge whether to trust a model for your use.
An example
Example: A hiring model trained only on a company's past hires can learn to repeat that company's past biases.
Related terms
Learn it properly
Training data is covered in AI Foundations, with short lessons, practice questions and instant feedback.