I’m not so sure, there are a lot of businesses and people training their AI models right now and sites like reddit or twitter are very attractive huge collections of user generated content. It’s not the most outrageous assumption that they’ll try to get that data for free by scraping instead of paying for API access.