designing a vote weighting system that favors similar instances

Would make the whole thing even worse, as I could create several new instances with 10 bot users each, then hammer out the votes.

The entire problem is that you can’t trace back each vote to a genuine user. It would be bad in case of fake instances that create 100 user accounts and upvote/downvote stuff, but you can ban the instance. It would be a disaster if a big instance creates fake votes (like lemmy.world suddenly adds 1000 fake users and uses them to manipulate other instances, if votes were anonymous you couldn’t check if it’s genuine lemmy.world users or fake accounts).

It would be damn easy to look up the instance and their “users” and see that the users are not genuine. Then ban the whole instance.

The problem isn’t keeping votes anonymous, that’s easy. The problem is bots/spam. You could just create a new instance and then upvote a post from another instance a thousand times. If the votes are anonymous for the other instance it’s tough to say if they are genuine users or just bots.

That’s the main issue here, when votes are anonymous you could easily just spam votes with no way to trace it back. If it’s a rogue instance then fine, you can ban the whole instance. But imagine if lemmy.world starts using fake votes in the background towards other instances.

Only humans can hold copyrights.

Yeah, no. Most copyrighted material is owned by companies, you don’t have to be a natural person to hold copyrights. And if a company can hold copyrights, you can also argue it can have fair use.

If I as a human listened to every single song of a band from start to finish, then produced a similar song in the same vein (lyrics / music genre), it would be fair use.

So why would it stop being fair use if an AI does the same thing? Just that the AI can listen to every song of this band and a million other bands, combining them.

But that’s not how model training works, it doesn’t simply copy and paste entire songs into its training data. It more or less “listens” to it, analyzes it and when you ask to create a rock song for example it just has an algorithm behind it what a song like that would sound like.

But you can’t just ask it to generate Bohemian Rhapsody from its data, it would probably get very close depending on the training, but it would never be 100% the same (except the model was only trained on this one song).

Just like you can listen to rock songs and then make your own, that’s totally valid. The problem here is of course automation and scale, but saying it’s not fair use is dubious.