Due to the American artificial intelligence research company OpenAI Collecting information from all over the Internet without authorization to train its AI chatbot, The right holders have filed lawsuits against him. This time it was brought up by the author, They call ChatGPT Violated the copyright of their novel.

The proposed class-action lawsuit filed in federal court in San Francisco says, OpenAI "Use a large number of copyrighted works, Without consent, Unmarked reference, There's no compensation" . The plaintiffs are asking the court to rule that the company violated the copyright of the authors' works by illegally downloading copies of novels to train its AI systems, ChatGPT The output of the answer constitutes infringement.

Generative AI companies have faced legal challenges for using materials to train their AI systems, The court is debating whether the practice qualifies as fair use. OpenAI Is facing a proposed class-action lawsuit, Accused its AI technology of analyzing billions of lines of computer code to generate its own qualified for copyright infringement. Getty Images is also suing the AI Art Generator Stable Diffusion Copyright infringement.

The lawsuit filed by the authors states, ChatGPT A summary of their novel was produced at the prompt, This is evidence of infringement. They think, this "Only in ChatGPT Abstract output is only possible if you are trained on the plaintiff's copyrighted work" .

The suit says, Because AI systems can't function without extracting information from materials, for ChatGPT The software program that powers it is called a large language model "Is itself a derivative work of infringement, Without the plaintiff's permission, Violated them in "Copyright law" Under exclusive rights" , A derivative work is a copyrighted work based on a pre-existing one.

Author pair OpenAI The illegal downloading of hundreds of thousands of books to train its AI systems has raised objections. 2018 years 6 month, OpenAI disclose, It goes towards GPT-1——The first iteration of its large language model——provided BookCorpus superior 7000 Several collections of stories, BookCorpus It was put together by an artificial intelligence research team.

The suit says: "They came from a group called Smashwords. com These books are reproduced on the website, The site hosts unpublished novels, Readers can read it for free. however, These novels are basically copyrighted. They were copied to BookCorpus Data set, Without the author's consent, Unmarked reference, There's no compensation. "

The lawsuit states that, OpenAI Later versions of the larger language model are trained on more copyrighted works. in 2020 Annual introduction GPT-3 In the paper, Disclosure of the company, Its training data set 15%Be from "2 An Internet-based book corpus" , It's simply called "Books1" and "Books2" . Although it has never revealed what works these datasets are, But the author claims they came from "The infamous Shadow library website" , Such as Library Genesis, Z-Library, Sci-Hub and Bibliotik.

Author's lawyer, Had represented programmers on OpenAI And Microsoft filed a class action lawsuit against Joseph．Saveri (Joseph Saveri) write: "These blatantly illegal shadow libraries have long been of concern to the AI training community. Such as, EleutherAI in 2020 years 12 The October release is called 'Books3' The AI training dataset has Bibliotik Reappearance of the collection, Include near 20 Ten thousand books. "

OpenAI Said last year, In view of GPT-4 Such as the competitive status and security implications of large-scale models, It no longer discloses information about the sources of its datasets.

The nationwide class-action lawsuit on behalf of hundreds of thousands of authors in the United States was filed by Paul．Tremblay (Paul Tremblay) And Mona．Awad (Mona Awad) mentioned. Tremblay wrote the novel "The cabin at the end of the world (The Cabin at the End of the World) " , The novel was written by Knight．Shyamalan (M. Night Shyamalan) Be adapted to "Visit the cottage (Knock at the Cabin) " . Indictment charge OpenAI Direct copyright infringement, Vicarious copyright infringement, violate "Digital Millennium Copyright Act" , Unjust enrichment and dereliction of duty and other torts.

OpenAI And Microsoft, which partly owns the artificial intelligence company, did not immediately respond.

Before the Judiciary Committee of the United States House of Representatives, Subcommittee on Intellectual Property and the Internet 5 At a hearing in October examining the intersection of artificial intelligence and copyright law, Major actors in Hollywood have advocated legislation against rampant, The unauthorized collection of works to train the behavior of artificial intelligence systems. ASHLEY, president of the Society of Composers and Lyricists．Irwin (Ashley Irwin) Said at the hearing: "Unless immediate measures are taken to address these emerging problems in terms of legal interpretation and economics, Otherwise, the rapid introduction of generative AI systems will be seen as a threat to creative careers. Policies and regulations must be prioritized, To protect the intellectual property rights and Copyrights of creators, And preserve the diverse and vibrant American cultural landscape. "

Owen emphasizes, Ai companies should be required to obtain the consent of creators when using their work to train AI programs, And compensate them at fair market value for any new work they subsequently create, Also annotate references. (Be compiled from www. hollywoodreporter. com)

TRANSLATORS: Rason group　proofread: Wang Dan

disclaimer: This network reprint or compile the original articles are from the network, Does not represent the views of this website or confirm the authenticity of its content. If the source is mislabeled or the copyright of the article is involved, Please contact us, This website will be corrected in due course, delete, thank you.