The group of authors who sued AI company Anthropic for copyright infringement announced a $1.5 billion settlement deal on Friday. That deal is now being discussed in the court where the authors filed their lawsuit.
Legal representatives for the authors said the “massive payout” secured from Anthropic “sends a powerful message of accountability to AI developers who torrented copyrighted works from illegal pirated websites to train AI models”. While also “giving hope to creators of every kind, including the writers, musicians, artists, journalists and others seeking to enforce creators’ rights in dozens of other pending cases”.
The authors said that Anthropic infringed their copyrights when it made unlicensed copies of their books in order to train its Claude AI model. Anthropic, like many AI companies, argued that AI training constitutes ‘fair use’ under American copyright law, and therefore it didn’t need permission or a licence from any authors or publishers when it made copies of their books.
The judge overseeing the case, William Alsup, actually agreed that AI training is fair use. But only if an AI company first secures legitimate copies of whatever works it is copying. Although Anthropic had bought and scanned some physical books, it also downloaded millions of pirated ebooks from piracy websites. Which meant it faced liability for copyright infringement in relation to all those pirated downloads.
Hence why Anthropic entered into settlement talks, despite in theory winning the fair use argument. Lawyers working for the authors reckon that about 500,000 books copied by Anthropic will be covered by this settlement. The AI company will pay around $3000 in damages per book.
The legal claims of any authors not covered by this settlement will not be affected by this deal, plus the settlement only relates to the copying of works on the input of the Claude AI. If it can be shown that Claude is infringing copyright on the output – by generating works that are the same as existing works – a separate legal claim could still be made.
Attorney Justin Nelson of Susman Godfrey LLP says: “This landmark settlement far surpasses any other known copyright recovery. It is the first of its kind in the AI era”.
“It will provide meaningful compensation for each work and sets a precedent requiring AI companies to pay copyright owners”, he adds. “This settlement sends a powerful message to AI companies and creators alike that taking copyrighted works from these pirate websites is wrong”.
Obviously creators and copyright owners – including in the music industry – argue that AI training is never fair use, and would have preferred it if the judge in this case had ruled that way.
Indeed, had the lawsuit gone to appeal, a higher court might have reached a different conclusion. And in the other AI copyright cases, including those filed by the music industry, creators and rightsholders will continue to argue against the fair use defence.
However, given how many AI companies are believed to have relied on illegitimate sources of content, Alsup’s ruling could help force those companies into speedy and lucrative settlements and licensing deals. Which is possibly a better option for creators and rightsholders rather than fighting long drawn out legal battles testing the fair use defence all the way to the US Supreme Court.
Of course, there might be some debate in the creator community as to how good a deal this settlement really is, given Anthropic recently closed a $13 billion funding round with a $183 billion valuation. And had the case proceeded through the courts, with millions of ebooks pirated and statutory damages in the US of up to $150,000 per infringement, the AI company could have faced a trillion dollars in damages.
But maybe a $1.5 billion pay day now is a good result. Especially with US tech companies continuing to push their fair use arguments, and with Donald Trump going around telling everyone that forcing American AI businesses to get a licence for everything they’ve copied to date just “isn’t doable”.
And even if the courts do ultimately side with the copyright owners, it’s going to take years to take the big test cases through all the appeals and onto the Supreme Court.
Among those welcoming the Anthropic settlement this weekend on the creator and rightsholder side was the Human Artistry Campaign, set up by the US music industry to lobby on AI issues on behalf of the wider creative industries.
It said in a statement, “this settlement is a huge victory not only for the authors involved, but for all writers, artists and creators who know that their work has value and their rights should be respected when it comes to AI uses”.
“We hope this is just the first of many AI companies to be held accountable for their theft of creative work”, it adds, “and that it helps set a precedent that consent and compensation for works used in AI training are nonnegotiable. When it comes to AI accountability and fighting for creators’ rights, this is just the beginning – but a powerful, historic start”.
