More authors sue OpenAI and Meta over copyright due to training

A fresh wave of authors, including renowned names like Michael Chabon and David Henry Hwang, have initiated class-action copyright infringement lawsuits against tech giants OpenAI and Meta. The writers allege that these AI pioneers employed unauthorised copies of their literary works in training their AI models, a claim that extends to copies purportedly sourced from notorious pirate websites. Filed in the Northern District of California just last week, these suits feature notable authors, including Matthew Klam, Rachel Louise Snyder, and Ayelet Waldman.

Read: Authors’ pirated books used to train Generative AI

According to the complaint filed against OpenAI, “On information and belief, the reason ChatGPT can generate writing in the style of a certain author or accurately summarize a certain copyrighted book and provide in-depth analysis of that book is because it was copied by OpenAI and copied and analyzed by the underlying GPT model as part of its training data,” The complaint further asserts that, “at no point did Plaintiffs authorize OpenAI to download and copy their protected works.”

Authors allege unauthorised copying

These allegations bear a striking resemblance to two earlier author-initiated class-action lawsuits, featuring Paul Tremblay and Mona Awad (though Awad has since withdrawn from the suit), as well as a separate suit brought forth by Christopher Golden, Richard Kadrey, and comedian Sarah Silverman. In the realm of class action law, it is commonplace for multiple suits to be filed independently and later consolidated into a single proceeding.

Read: AI open letter: authors including Margaret Atwood urge companies to honour copyright

Much like their predecessors, the most recent legal action seeks both damages and injunctive relief, and lists six causes of action: direct copyright infringement, vicarious copyright infringement, a DMCA violation for the removal of copyright management information, violations of California’s unfair competition law, negligence, and unjust enrichment. In the most recent lawsuit, Snyder said she had joined the cause because “the future of creatives matter.”

OpenAI’s defence and the challenging road ahead

While this legal battle unfolds, OpenAI has made efforts to contest the claims against them by the authors. In an August 28 filing, OpenAI’s legal team sought the dismissal of five of the six causes of action brought forth by authors, citing various deficiencies. OpenAI’s attorneys argue that, “None of these causes of action states a viable claim for relief because none of the legal theories challenged here actually condemns the conduct alleged with respect to ChatGPT, the language models that power it, or the process used to create them.” These motions are slated to be heard on December 7 as per a preliminary schedule.

The balance of fair use: navigating copyright in the age of AI creativity

Although these lawsuits have succeeded in drawing attention to the challenges posed by AI to creators, legal experts suggest that the copyright claims face an uphill battle in court. Even if these suits overcome the initial hurdles related to the alleged copyright infringements and the intricacies of AI training, lawyers argue that ample case law supports the notion of fair use.

They point to a recent case involving plagiarism detector TurnItIn.com, which upheld the idea that works could be ingested to create a database used to uncover student plagiarism. The landmark case of Kelly v. Arriba Soft also bolstered the argument for fair use by ruling that the reproduction and display of photos as thumbnails constituted fair use. In the realm of the publishing industry, the landmark Google Books case similarly underscores the transformative nature of using copyrighted material.

“I just don’t see how these cases have legs,” remarked one copyright lawyer bluntly in July, adding that there is “too much established case law to support this kind of transformative use as a fair use.” In a similar vein, federal judge William Orrick indicated his inclination to dismiss copyright claims filed against AI companies by artists in July, confirming the prevailing belief among legal and technology analysts that current copyright law struggles to address the complexities introduced by AI.

As this gripping legal drama unfolds, it remains to be seen whether these authors can carve out a legal precedent against OpenAI and Meta that reconciles the intricacies of AI-driven creativity with the established tenets of copyright law.