In a legal move, the Authors Guild, in collaboration with 17 distinguished authors, including George R R Martin, John Grisham, and Jodie Picoult, has initiated a class-action lawsuit against OpenAI, accusing the tech giant of copyright infringement. The suit, filed in the Southern District of New York, aims to address the unauthorized use of authors’ copyrighted works to train artificial intelligence models, specifically OpenAI’s GPT, and its far-reaching implications for the world of literature and creative content.
Authors defend their literary creations
Renowned authors and their legal representatives have united in a class-action lawsuit against OpenAI, alleging copyright infringement and raising significant concerns about the unlicensed use of books to train AI models. The named plaintiffs in the case, including David Baldacci, Mary Bly, Michael Connelly, Sylvia Day, Jonathan Franzen, John Grisham, Elin Hilderbrand, Christina Baker Kline, Maya Shanbhag Lang, Victor LaValle, George R R Martin, Jodi Picoult, Douglas Preston, Roxana Robinson, George Saunders, Scott Turow, and Rachel Vail, have come forward to protect their literary creations.
Rachel Geman, a partner with Lieff Cabraser and co-counsel for the plaintiffs, stressed the significant impact of copyrighted works on shaping AI technologies. She pointed out that the absence of the plaintiffs’ and the proposed class’ copyrighted works would have resulted in a substantially altered commercial product for the defendants. Geman highlighted that the defendants’ choice to replicate authors’ works without granting any alternatives or extending compensation posed a substantial threat to the role and livelihood of writers in its entirety.
Scott Sholder, a partner with Cowan, DeBaets, Abrahams & Sheppard and co-counsel for the plaintiffs, made it clear that the plaintiffs did not oppose the advancement of generative AI. But, he asserted that the defendants lacked the authority to develop their AI technologies by employing authors’ copyrighted works without permission. Sholder also suggested that the defendants had alternative options, such as training their large language models using works available in the public domain or compensating authors through reasonable licensing fees for the use of copyrighted materials.
The Authors Guild, a prominent advocate for writers’ rights, has taken a resolute stance against the unauthorized use of books to train AI models. They initiated this lawsuit after witnessing the harm and existential threat posed to the author profession by the unlicensed use of literary works to create large language models that generate text. The council of the Authors Guild and the board of the Authors Guild Foundation voted unanimously, with some abstentions, to file the lawsuit, driven by the profound unfairness and danger of using copyrighted books to develop commercial AI machines without permission or compensation.
Maya Shanbhag Lang, President of the Authors Guild, reiterated their unwavering commitment to safeguarding authors against what they view as theft by OpenAI and similar generative AI entities. She emphasized the guild’s mission to protect the literary world and the profession of writing. Lang highlighted their unique position as the oldest and largest writers’ association, boasting nearly 14,000 dedicated members. She also stressed the expertise of their formidable legal team in copyright law, emphasizing that the decision to initiate the lawsuit was a serious one, reflecting their resolute dedication to protecting authors’ rights.
AI’s impact on literature
The Authors Guild is deeply concerned about the potential damage to literary culture caused by the unauthorized use of authors’ works for AI training. They argue that AI models like GPT can produce content imitating human authors, resulting in derivative works based on copyrighted material. Mary Rasenberger, CEO of the Authors Guild, stressed the urgency of stopping this unauthorized usage to protect the thriving literary culture that underpins various creative industries in the US. She highlighted that great books are the product of authors dedicating their careers to honing their craft. Rasenberger emphasized the importance of authors retaining control over when and how their works are employed by generative AI to uphold the integrity of literature.
Authors contend that AI models, such as GPT, primarily create content that is a remix of existing works, lacking the unique human voice and creativity that define original literary works. They view the use of their characters and stories in AI-generated content as identity theft on a grand scale, as well as a violation of their intellectual property rights.
Franzen, one of the plaintiffs, highlighted the new and potentially exploitative nature of the generative AI field, especially in Silicon Valley. He argued that authors should have the autonomy to determine whether their works are utilized for AI training and, when they opt in, should receive fair compensation for their contributions.
AI lawsuit addressing concerns beyond fiction
The class-action suit focuses primarily on fiction writers as the first step, given the widespread mimicking of works of fiction by GPT and similar AI models. But, the Authors Guild acknowledges the harm extends to non-fiction markets as well. They are committed to addressing these concerns across various literary genres. A positive outcome of this case is expected to benefit writers from all backgrounds and genres, protecting their intellectual property rights and ensuring fair compensation for the use of their works in training AI.
This lawsuit comes in the wake of a similar class-action suit filed in July, which exposed that OpenAI’s ChatGPT and Google’s BERT were trained using over 7,000 books scraped from Smashwords, a platform for self-publishing e-books, without permission. In the same month, over 8,000 authors, including prominent names like Margaret Atwood, Jennifer Egan, George Saunders, and Jodi Picoult (now a plaintiff in the current lawsuit), signed an open letter led by the Authors Guild, urging Generative AI leaders to obtain consent, provide proper credit, and fairly compensate writers for the use of copyrighted materials in training AI models.