Book Authors’ ChatGPT Copyright Violation Claims Get Thrown Out By Judge

Image generated on AI

Its robots: 1, humans: 0 as a California district judge has recently made a page-turning decision leaning in favor of OpenAI amid allegations of copyright infringement by book authors. The lawsuit, initiated by a group of writers including well-known figures like Sarah Silverman, Michael Chabon, and Ta-Nehisi Coates, accused the artificial intelligence giant of training its sophisticated chatbot, ChatGPT, on pirated book copies without obtaining permission. They’ve labeled the technology a high-tech “grift.”

The authors insisted that ChatGPT’s ability to generate text based on their protected works constituted a violation of copyright laws, alongside claims of unfair business practices and unjust enrichment under state laws.

However, Judge Araceli Martínez-Olguín has dismissed most of these claims, particularly the accusation of the “direct” taking of a leaf out of the plaintiffs’ books, stating that they failed to demonstrate that ChatGPT’s responses are “substantially similar” or even “similar at all” to the original copyrighted materials, per Ars Technica.

Martínez-Olguín’s ruling emphasized that the authors did not provide evidence of ChatGPT producing outputs that prominently borrow from the copyrighted books. Without allegations of direct copying, the plaintiffs needed to prove substantial similarity between ChatGPT’s outputs and the content in question, a requirement they did not meet according to the judge.

Furthermore, the complaint’s claims under the Digital Millennium Copyright Act (DMCA) were also dismissed. The authors had accused OpenAI of intentionally removing copyright management information from its training data, a charge the judge found unsupported by sufficient facts.

However, the judge did allow some accusations to proceed under California’s unfair competition law, specifically regarding the unauthorized use of copyrighted works to train ChatGPT.

Martínez-Olguín also rejected allegations related to negligence and unjust enrichment, pointing out that the plaintiffs had only alleged intentional acts by OpenAI and failed to clarify how the company unjustly benefited from using their copyrighted works for ChatGPT’s training.

The authors now face a deadline of March 13 to consolidate their complaints and amend their arguments if they wish to pursue any of the dismissed claims further.

[via Ars Technica and Hollywood Reporter, cover image generated on AI]