In a high-stakes showdown that could shape the future of copyright protection in the era of generative AI, The New York Times is contemplating legal action against OpenAI. According to confidential sources, the newspaper giant is exploring the possibility of suing OpenAI to safeguard its intellectual property rights.
Tensions between The New York Times and OpenAI have been brewing for weeks as they negotiate a licensing deal. The deal would require OpenAI to compensate the Times for incorporating its news stories into the tech company’s AI tools. However, these discussions have deteriorated to the point where the Times is now contemplating legal action.
Sources familiar with the matter, who requested anonymity, revealed that lawyers for the newspaper are actively considering the option of legal recourse. A lawsuit from The New York Times against OpenAI could set the stage for one of the most high-profile legal disputes centered around copyright protection in the age of generative AI.
The fear of AI competition
One of the primary concerns driving the Times to consider legal action is the perception that ChatGPT, an AI model developed by OpenAI, is evolving into a direct competitor by generating text that answers questions based on the newspaper’s original reporting and writing. This fear is compounded by tech companies’ increasing use of generative AI tools, such as Microsoft, which has integrated ChatGPT into its Bing search engine.
The worry is that when users search online and receive AI-generated answers drawn from the Times’ reporting, it reduces their need to visit the newspaper’s website. This has raised significant concerns about the impact on the Times’ readership and revenue.
The legality of data collection
Generative AI models like ChatGPT gather vast amounts of data from the internet to inform their responses to various inquiries. This data collection often occurs without explicit permission, raising questions about its legality. If OpenAI is found to have violated copyrights during this data collection process, federal law permits the destruction of the infringing articles at the end of the legal proceedings.
Under federal copyright law, those found guilty of willful infringement may face substantial financial penalties, potentially reaching up to $150,000 for each infringement. This legal threat looms large over AI companies, emphasizing the need for negotiated solutions.
The Times’ negotiations with OpenAI come after reports that the newspaper has chosen not to join other media organizations in negotiating with tech companies regarding using their content in AI models. A representative of the Times declined to comment on whether this decision is related to the potential litigation against OpenAI.
However, the Times’ executives have publicly acknowledged the tension surrounding these issues. In June, Times CEO Meredith Kopit Levien emphasized the need for tech companies to compensate the newspaper for using its extensive archives. She argued that there should be a fair value exchange for past and future content used to train AI models.
Legal action on multiple fronts
If The New York Times decides to pursue legal action, it will join a growing list of copyright holders aiming at AI companies. Comedian Sarah Silverman has joined a class-action lawsuit against OpenAI, alleging unauthorized use of her memoir in ChatGPT. Other AI companies, such as Stability AI, have also faced copyright lawsuits, with Getty Images suing Stability AI for allegedly training an AI model on over 12 million photos without authorization.
Legal experts suggest that AI companies may invoke the “fair use doctrine” in their defense, which allows for using copyrighted material in specific instances, such as teaching, criticism, research, and news reporting.
The fair use doctrine and legal precedents
Two legal precedents are expected to be pivotal in upcoming AI copyright disputes. First, a 2015 federal appeals court ruling deemed Google’s digital scanning of millions of books for its Google Books library as a legally permissible use of “fair use,” not copyright infringement. The court’s decision was based on the determination that Google’s digital library did not create a significant market substitute for the original works.
Proving that ChatGPT and similar AI models do not pose a significant market substitute for the Times’ content will be a substantial challenge for OpenAI.
The second relevant case is the Supreme Court’s May decision in the Andy Warhol Foundation case. The court ruled that Andy Warhol’s alteration of a photograph of Prince was not protected by the fair use doctrine, particularly because both Warhol and the photographer were selling their images to magazines. This ruling highlights the importance of whether the original and copied works serve similar purposes or risk substitution.
Lawyers representing The New York Times argue that OpenAI’s use of the newspaper’s articles to generate news event descriptions should not be protected by fair use. They contend that it could potentially serve as a replacement for the Times’ coverage.
As these legal battles unfold, the future of copyright protection in the age of generative AI hangs in the balance. The outcome of these disputes will shape how AI companies, media organizations, and content creators navigate the complex landscape of intellectual property rights in the digital era.
