AI companies can just scrape the content, and since the sites don't have an exclusive license to it (it's user-generated, after all) there is little they can do

AI companies can just scrape the content, and since the sites don't have an exclusive license to it (it's user-generated, after all) there is little they can do against that. Actually, I'm not even sure the AI companies don't do this already. If I'd start such a company, I'd invest in a generic crawler once, instead of building countless integrations for all the different, proprietary APIs or dump formats.