Archive of Our Own writers are making their accounts private to prevent their fanfiction from being used to train AI models.
Article by TechCrunch: Writers on the fan fiction platform AO3 are increasingly locking access to their stories for non-registered users in order to avoid the crawlers of AI companies scrapping the text for training of their AI models.
So, if AO3 wanted they could make a few new tags, let people post generated content specifically for training data scrapers but posted as if it’s authentic to help advance model collapse. It would potentially make these companies see AO3 as a poison pool, especially if users change up the tag about it often to keep the devs confused.