Training data — Training data is the large body of text and other content an AI/SI model learns from before it is released. Web pages, listings, and documents often become training data. Facts that appear consistently across many sources are more likely to be learned correctly. Some website owners choose to block training crawlers to protect proprietary content.

Added October 21, 2026