Public posts only… Not private posts. Honestly, if you ever post anything publicly on the internet (on any site, including here on Lemmy), you should assume it’s going to be used for all sorts of things.

I have no doubt that Lemmy posts are part of the data being ingested into AI training. Lemmy and other Fediverse apps are extremely easy to scrape since anyone can subscribe to a Lemmy community, Mastodon account, etc with ActivityPub and get all new posts and comments pushed to them. Don’t even have to actually scrape anything.