Training crawler
A training crawler is an AI crawler that gathers web content to train future models, rather than to answer a live question. OpenAI's GPTBot is the best-known example.
What this means for your store
Blocking training crawlers keeps your content out of model training, which some brands choose for control or copyright reasons. Important nuance: it does not remove you from live AI answers, because those are fetched by a different, retrieval crawler. Blocking training and blocking search are two separate decisions.