I mean, how is a bot not already crawling through public sites like Lemmy and Mastodon for the purposes of AI training? Federated or unfederated, if you are providing social media services that data you have is already out there.