Datasets

  • Release version: Australia
  • Updated March 12, 2026
  • 1 minute to read

    • Explore the datasets in detail with examples.

    A dataset is a collection of data used to train or evaluate an AI model. You can think of it as a large spreadsheet with rows representing examples and columns for features or labels. Datasets can also come in other formats, such as images, audio, or video. They help models learn patterns and decide.

    Mapping datasets within the AI Control Tower is important for tracking data sources, ownership, and other key details, making it a vital part of the organization's AI governance program.

    Examples:

    • A set of labeled images (dogs, cats) for training a model
    • Articles and books for training a language model
    • Customer purchase history for recommendations
    • Medical records for diagnosis tools