A federal judge in New York has ruled that OpenAI must face a consolidated multi-district class action filed by dozens of authors alleging direct copyright infringement by outputs of its large language model, ChatGPT. U.S. District Judge Sidney Stein of the Southern District of New York rejected OpenAI’s attempt to dismiss the claims, stating that “a more discerning observer could reasonably conclude that the allegedly infringing outputs are substantially similar to plaintiffs’ copyrighted works.”
The court applied the “more discerning observer” test, which focuses on whether the protected elements, such as plot, setting, and characters of a copyrighted work, have been reproduced. Judge Stein noted that even under the less precise “ordinary observer” test, the result would have been the same.
The decision highlighted examples involving summaries of George R.R. Martin’s A Game of Thrones series that OpenAI’s ChatGPT allegedly generated. Judge Stein replicated several outputs and found that “the summary conveys the overall tone and feel of the original work by parroting the plot, characters, and themes of the original.” He concluded that a reasonable jury could find substantial similarity between ChatGPT’s outputs and Martin’s original works. The court, however, clarified that it has not yet ruled on whether the outputs are protected under the doctrine of fair use.
