So the problem regarding AI is more nuanced and complicated than the plain copyright-question of piracy. It's more akin to cases of plagiarism, which go case by case.
There at least two documented cases of the major AI companies downloading millions of books of torrents. Anthropic is in litigation about it right now, meta was in the news about it. I would be surprised if it's not all of them.
Very few people download(ed?) movies or music with the intent to distribute it. They downloaded it because they wanted to consume the media.
Come again? Deepseek and numerous other smaller models can be run locally, for free.
I think what’s really behind this attitude is that “information wants to be free” when I’m pirating IP from a big, faceless corporation. But when it’s _my_ IP being pirated, it’s theft.
In some developed nations, we're almost comically unaware of power dynamics, sometimes to the point of inverting power dynamics.
Consider a McDonald's employee versus a Police Officer. The McDonald's employee can make almost any mistake and it's not a big deal. Pickles on your burger when you didn't want pickles? Eh, you'll get over it.
When Police Officers make mistakes, innocent people die. Taxpayers pay out sometimes millions of dollars.
Based on that, you would expect the standard of behavior for Police Officers to be much higher than a McDonald's cashier. But no, we've inverted it.
If the Police Officer accidentally kills someone they shouldn't, they'll likely keep their job. If a McDonald's cashier's drawer comes up 1 dollar short, they're terminated immediately.
Oops. Messed that up, didn't we?
I’d like to point out that for decades the argument pro-piracy folks made was that you can’t “steal” software since it still remains. It’s only fair to apply that same standard to AI companies who are simply scraping data...