An image of an archeologist adventurer who wears a hat and uses a bullwhip

I was really hoping that the conversation around AI art would at least be partially centered on the perhaps now dated "2008 pirate party" idea that intellectual property, the royalty system, the draconian copyright laws that we have today are deeply silly, rooted in a fiction, and used over and over again, primarily by the rich and powerful, to stifle original ideas and hold back cultural innovation.

Unfortunately, it's just the opposite. It seems most people have fully assimilated the idea that information itself must be entirely subsumed into an oppressive, proprietary, commercial apparatus. That Disney Corp can prevent you from viewing some collection of pixels, because THEY own it, and they know better than you do about the culture and communication that you are and are not allowed to experience.

It's just baffling. If they could, Disney would scan your brain to charge you a nickel every time you thought of Mickey Mouse.

I don't really care.

Either enforce the current copyright regime and sue the AI companies to dust.

Or abolish copyright and let us all go hog wild.

But this halfway house where you can ignore the law as long as you've got enough money is disgusting.

This may not be a particularly popular opinion, but current copyright laws in the US are pretty clearly in favor of training an AI as a transformative act, and covered by fair use. (I did confirm this belief in conversation with an IP attorney earlier this week, by the way, though I myself am not a lawyer.)

The best-positioned lawsuits to win, like NYTimes vs. OpenAI/MS, is actually based on violating terms of use, rather than infringing at training time.

Emitting works that violate copyright is certainly possible, but you could argue that the additional entropy required to pass into the model (the text prompt, or the random seed in a diffusion model) is necessary for the infringement. Regardless, the current law would suggest that the infringing action happens at inference time, not training.

I'm not making a claim that the copyright should work that way, merely that it does today.