<\/p>\n
\n\tThe leading trade group for authors has entered a legal battle against OpenAI over its human-mimicking chatbot in a case that could decide the legality of using copyrighted works to train AI systems.<\/p>\n
\n\tThe Authors Guild \u2014 led by prominent fiction authors, including George R.R. Martin, Jonathan Franzen and John Grisham \u2014 on Tuesday sued OpenAI, accusing the company of engaging \u201cin a systematic course of mass-scale copyright infringement\u201d to \u201cpower their lucrative commercial endeavor.\u201d The proposed class action filed in New York federal court builds upon arguments from creators, who have already initiated lawsuits against AI firms, that generative AI illegally produces infringing works that directly compete with their creations. <\/p>\n
\n\tThe legal action is at least the third against OpenAI over the company using copyrighted books to train its system. OpenAI is facing a proposed class action from author Paul Tremblay, in addition to another filed by Sarah Silverman, which also names Meta. Artists have similarly sued AI art generators Stability AI, Midjourney and DeviantArt for copyright infringement.<\/p>\n
\n\tOpenAI has said that it trains its model on \u201clarge, publicly available datasets that include copyrighted works.\u201d While it remains unknown which works are included in the dataset, the authors point to ChatGPT generating summaries and in-depth analyses of the themes in their novels as proof that OpenAI used their books.<\/p>\n
\n\t\u201cFor example, when prompted, ChatGPT accurately generated summaries of several of the Franzen Infringed Works, including summaries for The Corrections<\/em>, Purity<\/em>, and Freedom<\/em>,\u201d states the complaint.<\/p>\n \n\tThese summaries, the authors say, constitute infringing works that interfere with their economic prospects to profit off of their books. They take issue with OpenAI choosing not to license their novels.<\/p>\n \n\tAI companies thus far have resisted striking licensing deals with authors, opting to obtain the data from internet-based book collections. Authors have argued in several lawsuits that the datasets were obtained from illegal shadow library sites, including Library Genesis, Z-Library and Bibliotik.<\/p>\n \n\tStressing that ChatGPT undermines the market for authors\u2019 works, the lawsuit says that the tool facilitates the creation of infringing derivative productions. It cites businesses that sell prompts allowing users to create what are essentially works of fan fiction. When prompted, ChatGPT generated the \u201cnext purported installment of While the Patient Slept<\/em>, one of the Authors Guild Infringed Works, and titled the infringing and unauthorized derivative \u2018Shadows Over Federie House,’\u201d according to the complaint. A derivative is a work based upon a preexisting, copyright-protected work.<\/p>\n \n\t\u201cFor example, a business called Socialdraft offers long prompts that lead ChatGPT to engage in \u2018conversations\u2019 with popular fiction authors like Plaintiff Grisham, Plaintiff Martin, Margaret Atwood, Dan Brown, and others about their works, as well as prompts that promise to help customers \u2018Craft Bestselling Books with AI,’\u201d the lawsuit states.<\/p>\n \n\tWorks of fan fiction are considered infringing creations. Authors, however, typically don\u2019t take action unless the works are monetized.<\/p>\n \n\t\u201cThe prompting and output of AI that creates a summary or a new storyline using the same characters or the same themes would be very possibly be considered a derivative and infringing work,\u201d says Ed Klaris, an intellectual property lawyer and professor at Columbia Law School.<\/p>\n \n\tCourts could look to A&M Records v. Napster<\/em> for direction on whether OpenAI is facilitating the infringement of authors\u2019 novels through ChatGPT. A federal judge in that case found the peer-to-peer sharing service liable for contributory and vicarious copyright infringement, rejecting a fair use defense. The court found that Napster had a duty to control, or at least restrict, users infringing on artists\u2019 copyrights through digital audio file distribution.<\/p>\n \n\tThe courts will also wrestle with two Supreme Court cases that legal experts say could decide the issue of whether OpenAI can avail itself of a fair use defense. AI companies will likely point to precedent greenlighting the copying of works to generate noninfringing text responses from when the Authors Guild in 2005 sued Google for digitizing millions of books to create a search function for the works. A federal judge in that case rejected copyright infringement claims, finding the company\u2019s copying of the books to amounts to fair use. Central to the ruling was that Google only allowed users to view snippets of text without providing the full book.<\/p>\n \n\tThe authors, meanwhile, will likely cite a recent decision from the Supreme Court in Andy Warhol Foundation for the Visual Arts v. Goldsmith<\/em>. In that case, the justices effectively reined in the scope of fair use. The justices stressed that the defense is likely to be rejected when an original work and derivative share the \u201csame or highly similar purpose,\u201d and they\u2019re competing in the same market.<\/p>\n \n\tKlaris says the courts will likely reject fair use based on the Supreme Court\u2019s decision in Warhol. \u201cThis is a huge intellectual property grab that the AI companies are doing,\u201d he notes. \u201cIt\u2019s going to get reined in.\u201d<\/p>\n \n\tOpenAI in August moved to dismiss most claims in the copyright lawsuit led by Silverman. It argued that copyright law \u201cdoes not protect ideas, fact or language,\u201d according to the filing.<\/p>\n \n\t\u201cCopyright protects the particular way an author expresses an idea\u2014not the underlying idea itself, facts embodied within the author\u2019s articulated message, or other building blocks of creative expression,\u201d it stated.<\/p>\n