AI company Anthropic agrees to pay $1.5B to settle lawsuit with authors

a year ago by MicroWave to c/news

If approved, the settlement would be the largest in the history of American copyright cases, according to a lawyer for the authors behind the lawsuit.

Anthropic, a major artificial intelligence company, has agreed to pay at least $1.5 billion to settle a copyright infringement lawsuit filed by a group of authors who alleged the platform had illegally used pirated copies of their books to train large-language models, according to court documents.

“If approved, this landmark settlement will be the largest publicly reported copyright recovery in history, larger than any other copyright class action settlement or any individual copyright case litigated to final judgment,” said Justin Nelson, a lawyer for the authors.

The lawsuit, filed in federal court in California last year, centered on roughly 500,000 published works. The proposed settlement amounts to a gross recovery of $3,000 per work, Nelson said in a memorandum to the judge in the case.

frongt 44 points a year ago

It's a start. Now do openai, Facebook, and the rest of them. Hopefully it's not reduced on appeal.

save

path: 0 19231634, hotness: undefined, score: 44, children: 1

halcyoncmdr 19 points a year ago

Hopefully it's not reduced on appeal.

It's a settlement, not a judgement. Both sides agreed to this amount to end the lawsuit. Means there is no legal precedent set either.

save

path: 0 19231634 19232948, hotness: undefined, score: 19, children: 0

WalnutLum 28 points a year ago

Their predicted 2025 revenue is supposedly 5 billion so this is a decent chunk.

Not enough, but a decent chunk.

save

path: 0 19232047, hotness: undefined, score: 28, children: 12

FlowVoid 17 points a year ago

They were also forced to delete the pirated works from their training data. Which means they might have problems training the next version of their LLM, and if so that revenue is going to dry up.

save

path: 0 19232047 19233303, hotness: undefined, score: 17, children: 11

ArrantKnave 7 points a year ago

If true, great, but I fear they'll "delete" it.

save

path: 0 19232047 19233303 19233436, hotness: undefined, score: 7, children: 1

FlowVoid 3 points a year ago

I doubt it. How would they keep disgruntled ex-employees from spilling the beans?

save

path: 0 19232047 19233303 19233436 19234036, hotness: undefined, score: 3, children: 0

GissaMittJobb 3 points a year ago

They essentially still have the information in the weights, so I guess they won't fret too much over not having it in the original training data.

save

path: 0 19232047 19233303 19236048, hotness: undefined, score: 3, children: 8

FlowVoid 3 points a year ago

They will need actual training data when they want to develop the next version of their LLM.

save

path: 0 19232047 19233303 19236048 19237883, hotness: undefined, score: 3, children: 7

GissaMittJobb 3 points a year ago

I guess it depends on how important old data is when building upon new models, which I fully admit I don't know the answer to. As I understand it though, new models are not trained fully from scratch, but instead are a continuation of the older model trained with new techniques/new data.

To speculate, I guess not having the older data present in the new training stages might make the attributes of that data be less pronounced in the new output model.

Maybe they could cheat the system by trying to distill that data out of the older models and put that into the training data, but I guess the risk of model collapse is not-insignificant there

Again, limited understanding here, take everything I speculate with a grain of salt

save

path: 0 19232047 19233303 19236048 19237883 19239809, hotness: undefined, score: 3, children: 6

TooManyGames 19 points a year ago

This is interesting, because a judge just let Meta off the hook for downloading copyrighted works from pirate sites and even using Metas vast infrastructure to try to hide it. It's curious that Anthropic has to pay and Meta doesn't.

save

path: 0 19234843, hotness: undefined, score: 19, children: 2

pedroapero 6 points a year ago

That's a settlement, they did not have to, but voluntarily did before the verdict.

save

path: 0 19234843 19245249, hotness: undefined, score: 6, children: 1

TooManyGames 5 points a year ago

That clarifies it, so Anthropic kinda chickened out and considered it safer to make this a cost of doing business thing.

save

path: 0 19234843 19245249 19245689, hotness: undefined, score: 5, children: 0

ZoteTheMighty 8 points a year ago

Amazing, let's not forget that not a single AI company is profitable, and it's very possible that AI won't get substantially better than this. They're just shoveling billions of dollars into this thing as "the cost of doing business", and they probably won't see a real ROI from it.

save

path: 0 19233996, hotness: undefined, score: 8, children: 0

Arcane2077 6 points a year ago

Damn shame. America is a joke

save

path: 0 19234347, hotness: undefined, score: 6, children: 0