Anthropic judge rejects $1.5B AI copyright settlement

I'm an author, and I've confirmed that 3 of my books are in the 500K dataset.

Thus, I stand to receive about $9,000 as a result of this settlement.

I think that's fair, considering that two of those books received advances under $20K and never earned out. Also, while I'm sure that Anthropic has benefited from training its models on this dataset, that doesn't necessarily mean that those models are a lasting asset.

Be careful what you wish for.

While I'm sure it feels good and validating to have this called copyright infringement, and be compensated, it's a mixed blessing at best. Remember, this also means that your works will owe compensation to anyone you "trained" off of. Once we accept that simply "learning from previous copyrighted works to make new ones" is "infringement", then the onus is on you to establish a clean creation chain, because you'll be vulnerable to the exact same argument, and you will owe compensation to anyone whose work you looked at in learning your craft.

This point was made earlier in this blog post:

https://blog.giovanh.com/blog/2025/04/03/why-training-ai-can...

HN discussion of the post: https://news.ycombinator.com/item?id=43663941

It's a good thing that laws can be different for AI training and human consumption. And I think the blog post you linked makes that argument, too, so I'm not sure why you'd contort it into the idea that humans will be compelled to attribute/license information that has inspired them when creating art.

Right — laws can be arbitrary, and ignore constraints like consistency! It’s just something sane people try to avoid.

The inconsistency you're talking about is only based on the premise that LLMs and humans are "basically the same thing and thus should be treated the exact same way in this kind of situation". But I don't really see why that would be the case in the first place.

Now don't take me wrong, I'm not saying that a rushed regulatory response is a good thing, it's more about the delivery of your reply. I see those arguments a lot: people smugly saying "Well, YOU too learn from things, how about that? Not so different from the machine huh?" and then continuing the discussion based on that premise, as if we were supposed to accept it as a fact.

Because the onus is on you to show the substantive difference. Learning from copyrighted works has always been accepted as free and unrestricted, up until 2022. Before that, nobody (to a rounding error) thought that simply being influenced by a previous copyright work meant you owed a license fee. If anything, people would have been livid about restrictions on their right to learn.

Only when big-corp critics needed another pretense to support a conclusion they long agreed with for other reasons, did they decide that the right to learn from your exposure to a copyright work was infringement.

If you're interested in the similarities and genuinely curious, you could look at the article linked above, which shows how both LLMs and humans store a high-level understanding of their training set. It's a way deeper parallel than "it's all learning" -- but you have to be willing to engage with the other side rather than just strawman it.