OpenAI has urged a federal judge to reject parts of a lawsuit filed by the New York Times, claiming that the newspaper manipulated its ChatGPT chatbot and other AI technologies to fabricate evidence for their legal action. In a recent submission to a federal court in Manhattan, OpenAI accused the Times of employing deceptive tactics that contravene the company’s usage policies, leading to unauthorized reproduction of its content.
The company criticized the Times for not adhering to its high journalistic standards within the lawsuit’s claims, stating, “The real story, which will be revealed during the litigation, involves the Times compensating an individual to compromise OpenAI’s products.”
Although OpenAI refrained from identifying the external party allegedly engaged by the Times or from alleging any violation of anti-hacking laws, the situation has escalated tensions between the two entities.
Inquiries directed at representatives from both the New York Times and OpenAI regarding the court filing have yet to receive responses.
The lawsuit, initiated by the Times against OpenAI and its principal investor Microsoft in December, centers on the accusation of unlawfully training chatbots with millions of the newspaper’s articles to deliver content to users without proper authorization. This legal challenge is part of a broader wave of copyright disputes targeting technology companies over the alleged improper use of copyrighted content in AI development. These companies defend their practices by arguing that AI’s utilization of copyrighted materials falls under fair use, a stance they believe is crucial for the continued expansion of this burgeoning sector.
Despite these defenses, the core issue of whether AI training constitutes fair use under copyright law remains unresolved in the courts. To date, some claims of infringement related to the output of generative AI systems have been dismissed due to insufficient proof that AI-generated content closely mirrors copyrighted materials.
The complaint from the Times highlighted several instances where chatbots from OpenAI and Microsoft reproduced snippets of its articles almost exactly upon request, accusing them of leveraging the newspaper’s extensive journalistic efforts without permission.
OpenAI countered in its legal documents that achieving such specific outcomes from ChatGPT required “tens of thousands of attempts,” emphasizing the challenges in directly accessing Times articles through the chatbot under normal use.
Furthermore, OpenAI expressed confidence in the dismissal of the lawsuit based on the principle of fair use, arguing that AI’s learning process about factual information cannot be inhibited any more than one news outlet can block another from reporting on stories it did not originally investigate.