Another month, another copyright infringement suit filed against OpenAI.

In addition to previous litigation brought against artificial intelligence firms by the New York Times Company,  an alliance of prominent authors, and a group of creative artists, eight newspapers filed a complaint in district court in New York late last month, alleging that OpenAI and Microsoft are infringing their copyrighted articles by training generative AI products on their content and by churning out text that too closely resembles the copyrighted works.

And just like in the predecessor suits, the current litigation highlights a fundamental divide over AI that we’ve explored in this space on numerous occasions: While the newspapers regard ChatGPT and its ilk as mere automatons that mindlessly perform whatever operations they’re programmed to perform, OpenAI and Microsoft present their technology as genuinely autonomous (i.e. transformative and capable of transcending their rote programming.)

Mary Trimble at The Dispatch ably covered the details of the case, but in a nutshell, the eight newspapers—the New York Daily News, the Chicago Tribune, the Orlando Sentinel, the Sun Sentinel, the Mercury News in California, the Denver Post, the Orange County Register, and the St. Paul Pioneer Press—asserted that “Microsoft and OpenAI simply take the work product of reporters, journalists, editorial writers, editors and others who contribute to the work of local newspapers—all without any regard for the efforts, much less the legal rights, of those who create and publish the news on which local communities rely.” According to the publishers, not only ChatGPT but also Microsoft’s Copilot, which the software giant has integrated into its Windows and Office products, and its Browse with Bing offering, slavishly copies and pastes their copyrighted content into their training models without compensating them. 

In addition, the plaintiffs accused the defendants of “using the Publishers’ journalism to create GenAI products that undermine the Publishers’ core businesses by retransmitting ‘their content’—in some cases verbatim from the Publishers’ paywalled websites—to their readers.” They included in their complaint a number of instances where their prompts induced ChatGPT to spit out nearly identical versions of published articles. (Interestingly, the prompts were, essentially, “Please tell me about [Article X]. Please format your response as summary first, followed by the actual text.”) This is automatoner terminology: verbatim taking, copying, and retransmitting of existing content.

In response, an OpenAI spokesperson told the New York Times that the company was unaware of the publishers’ gripes before they went to court and reiterated that, “Along with our news partners, we see immense potential for A.I. tools like ChatGPT to deepen publishers’ relationships with readers and enhance the news experience.”

But previously, OpenAI has responded to litigation filed against it by describing ChatGPT as “a revolutionary technology with the potential to augment human capabilities, fostering our own productivity and efficiency” and praised generative AI as “a creative tool that can write sonnets, limericks, and haikus.” (Emphases added.) It has also contended that, as part of its training process, ChatGPT comes “to understand the facts that constitute humans’ collective knowledge” and employs “copyrighted content as part of a technological process that (as here) results in the creation of new, different, and innovative products.” (Emphases added.) These are textbook autonomist terms: understanding, creation, and augmenting human abilities—in essence, transcending the previous capabilities of computer technology.

Which model will win out in these cases remains to be seen. But it’s telling that, the day before the publishers filed suit, OpenAI announced a new partnership with the Financial Times¸ insisting, in classic autonomist language, that “our partnership and ongoing dialogue with the FT is about finding creative and productive ways for AI to empower news organizations and journalists, and enrich the ChatGPT experience with real-time, world-class journalism for millions of people around the world.” (Emphases added.)

Learn more: OpenAI Strikes Back at New York Times in Copyright Spat, Deepening the Philosophical Dispute | OpenAI-New York Times Copyright Fight Further Illustrates Autonomy-Automaton Dichotomy | Generative AI’s Napster Moment | Where Does Patent Reform Stand as 2023 Comes to a Close?