Part of:GenAI in court: AI training data copyright lawsuits to watch
The importance of the 'New York Times' AI copyright lawsuit
The newspaper publisher is the first major news outlet to sue the AI creator. While the suit might not reach court, it still has a significant impact on the AI community.
Listen to this article. This audio was generated by AI.
The list of organizations suing creators of generative AI systems over copyrighted material continues to grow.
The New York Times revealed on Dec. 27 that it has sued AI vendor OpenAI and its investor Microsoft for copyright infringement.
The Times contends that the vendors used millions of its articles to train automated chatbots that now compete with it as reliable sources of information. The newspaper alleges that the AI vendors should be responsible for "billions of dollars in statutory and actual damages" and wants them to destroy any chatbot models and training data that use copyrighted material from the Times.
While The New York Times is the first major news publisher to sue the AI creators, other copyright lawsuits have been filed in recent months.
In January 2023, Getty Images filed a complaint against AI image generator vendor Stability AI for copying and processing millions of its images and associated metadata.
Meta and OpenAI are facing copyright infringement suits from comedian, actor and writer Sarah Silverman. Silverman alleges that the vendors trained their large language models on her books. A federal judge dismissed most of her suit in late November.
Authors Paul Tremblay and Mona Awad also sued OpenAI for allegedly copying their books in its training data.
The importance of the NYT suit
The New York Times suit is similar to previous AI copyright suits, said Katie Gardner, an intellectual property lawyer with the firm Gunderson Dettmer.
"They're saying it's not fair use or transformative [a new work that uses an old work in new or unexpected ways] because there are many examples where they were able to prompt the models to get exact replicas of certain articles," Gardner said.
What's interesting, she added, is that the Times lawsuit stemmed from a failed negotiation.
The Times approached OpenAI in April to discuss concerns about the use of its copyrighted material and find a solution. However, that did not lead to an agreement.
Microsoft and OpenAI did not respond to requests for comment.
The New York Times is not the only publisher to seek a partnership agreement with OpenAI. In July, The Associated Press negotiated an agreement with the AI vendor to license the AP's archive of news stories. In exchange, the AP will get access to OpenAI's technology and product expertise, according to the two companies.
The lawsuit is also interesting because it is the first instance of a claim against a generative AI company with a significant public valuation, said Michael Bennett, director of education curriculum and business lead for responsible AI at Northeastern University.
The very high valuation of OpenAI is probably opening the eyes of several other companies ... that have been trying to figure out how to extract some value from their perception of having been exploited in this way for copyright-protected content.
Michael BennettDirector of education curriculum and business lead for responsible AI, Northeastern University
OpenAI is valued at more than $80 billion.
"The very high valuation of OpenAI is probably opening the eyes of several other companies ... that have been trying to figure out how to extract some value from their perception of having been exploited in this way for copyright-protected content," Bennett said.
Thus, if the Times can show ways that OpenAI infringed on its copyrighted material, it could ask for significant compensation.
The likelihood of settlement
However, it is unclear whether the case will end up being settled by both companies or if it will go to court.
While the continuation of the lawsuit in court could lead to a much-needed decision on copyright protection and fair use of AI technology and data, the process might be too complex and financially inconvenient for OpenAI, Bennett said.
"The likeliest outcome is that you see a settlement and see this case disappear so that you don't get a ruling at all, because there's too much uncertainty around that, and it's probably going to be more financially feasible from OpenAI and Microsoft's perspective," he said.
Moreover, it's also likely that The New York Times will follow Getty Images' strategy to partner with a significant AI vendor and use its journalistic content to train up models or build an AI application, Bennett added.
In September, Getty partnered with AI hardware and software provider Nvidia to launch a new AI-powered image generator.
Whether the parties settle or not, a major publisher like The New York Times suing a vendor like OpenAI and a tech giant like Microsoft could bring some clarification to the copyright issue, said Tinglong Dai, professor of operations management and business analytics at Johns Hopkins University's Carey Business School.
"It's going to help clarify legal thinking in terms of copyright and implications of generative AI," Dai said. While there is already a debate about copyright issues with AI models, having a big publisher such as The New York Times bring forth a suit will influence such discussions within the legal and AI research communities.
"Before the New York Times suit, people talked about it, but not really seriously," Dai said. But the publisher's suit is "one of the most compelling legal arguments against the ways generative AI is trained," he added.
Not only will it influence the AI research community, but it will also likely prompt vendors such as OpenAI to tweak their models and training data so that they don't infringe on the data of other companies, Dai said.
Esther Ajao is a TechTarget Editorial news writer covering artificial intelligence software and systems.
