Humanity said Friday in federal court in San Francisco that it agreed to pay $1.5 billion to resolve class actions by authors who claim that AI companies will use books to train Claude Chatbots without permission.
The mankind and the plaintiffs filed the matter asked Judge William Alsup to approve the settlement after announcing it in August without disclosing the terms or amounts.
If approved, this historic settlement will be the largest publicly disclosed copyright settlement in history.
The essence of reconciliation and its importance to the industry
The proposed settlement marked the first example of a resolution in a wave of litigation against high-tech companies including Openai, Microsoft and Meta platforms that used protected materials to train generation AI systems.
Authors Andrea Burtz, Charles Graver and Kirk Wallace Johnson filed a lawsuit against humanity last year. They argue that Amazon and Alphabet-backed companies illegally use millions of pirated copies to train assistant Claude to provide a more accurate response to user queries.
The author's allegations reflect dozens of other lawsuits that authors, media outlets, visual artists and high-tech companies claimed they had stolen their works to train artificial intelligence.
The company says it is engaged in the fair use of protected materials to create new, transformative content.
Alsup in June ruled that humanity was engaged in the fair use of the author's work to train Claude, but found that the company violated its rights by storing more than 7 million pirated books in the Central Library.
The trial is scheduled to begin in December to determine the amount of racial damages that are suspected of copyright infringement, and potential damages could range in the hundreds of billions of dollars.
The central issue of fair use continues to be debated in other copyright cases in the AI field. Another San Francisco judge said shortly after the Allsup ruling that using copyrighted work without permission to train AI is illegal in many circumstances, considering a similar case against Meta.
It is often illegal to use copyrighted materials without permission to train AI.
Therefore, this case emphasizes pressure on AI developers, clearly defining how copyright can be used for training without infringing authors' rights, and clearly defining what amounts to be set as a precedent for future cases.
