This is why we built nOne, a permission/consent layer underlying the engines/platforms/algorithms/etc.
This way OpenAI can check the permissions of the author of a particular body-of-work and then respect those permissions, so that the legality of their functionality/crawling is not questioned.
