That seems like a potentially very useful addition to the robots.txt "standard": Crawler categories. Wanting to disallow LLM training (or optionally only that

That seems like a potentially very useful addition to the robots.txt "standard": Crawler categories. Wanting to disallow LLM training (or optionally only that of closed-weight models), but encouraging search indexing or even LLM retrieval in response to user queries, seems popular enough.