Training Data
The collection of examples a model learns from — billions of images, texts, video clips, or songs, usually scraped or licensed at internet scale. Training data defines what a model can do: styles it saw a lot, it reproduces well; things it rarely saw, it fumbles. It is also the center of AI's biggest legal and ethical debates around copyright and consent. Practical takeaway for creators: if a model struggles with a niche concept, it likely had little training data for it — try describing the concept using more common visual terms instead.