Fediverse

Copyright is a law. Everything is copyrighted, with or without the little (C). Licensing is a peer-to-peer contract. Unless you can prove the other side is aware of and agreed to a contract, it doesn't bind them.

Notably, licensing often is needed because general copyright exists. The license grants them the right to copy your full text or whatever, and if they didn't agree to it, then they had no right to copy it. There are exceptions for excerpts and search indexing and the like, but they can't (legally) just take all your posts because you put them online.

That all said, big companies have already been doing mass copyright violations for AI, so copyright or licenses don't necessarily mean anything unless you can force them to comply. There are lawsuits on AI scraping now. Because the end result is either making up some reason that copyright doesn't ban copying if you do enough of it or making LLMs effectively illegal and putting some massive corporations on the hook for mass violations against basically everyone online, I wouldn't personally bet the courts ruling against the corporations.

It looks like I was mixing up some facts. The Genius case was denied because genius doesn't own the copyright to the lyrics they were publishing. I can't find the case now, but there was a case where a judge said scraping was allowed because it wasn't a given that the scraper had read a ToS.