OpenAI has filed a motion with a federal judge to dismiss a portion of a copyright infringement lawsuit brought against it by The New York Times, alleging that the newspaper paid individuals to manipulate OpenAI's artificial intelligence systems in order to produce misleading evidence for the case.
In a filing submitted to the Manhattan federal court on Monday, OpenAI asserted that The New York Times induced its technology to replicate copyrighted material through deceptive prompts that violated OpenAI's terms of use. However, OpenAI has refrained from revealing the identities of the individuals it claims were hired by The New York Times for these purposes, in order to avoid potential accusations of violating anti-hacking laws.
The document from OpenAI stated that the allegations presented in The New York Times' complaint fail to meet the high journalistic standards for which the publication is renowned. It further asserted that the evidence would reveal that The New York Times engaged in the practice of paying individuals to hack into OpenAI's products.
According to Ian Crosby, legal counsel for The New York Times, OpenAI's allegations of hacking were merely attempts to utilize OpenAI's products to gather evidence of alleged theft and replication of copyrighted material from The New York Times. In December 2023, The New York Times initiated a lawsuit against OpenAI and its primary financial supporter, Microsoft, accusing them of unauthorized use of millions of articles from The New York Times to train chatbots providing information to users. This lawsuit draws upon constitutional and copyright law to defend the original journalism of The New York Times, and also alleges that Microsoft's Bing AI system directly excerpts content from The New York Times.
The case is part of a broader trend where copyright holders, including writers, visual artists, and music publishers, are taking legal action against technology companies for allegedly misappropriating their content in the training of artificial intelligence systems. OpenAI contends that training advanced AI models without incorporating copyrighted material is infeasible, given the expansive scope of copyright protection. While the legal landscape regarding the fair use of AI training data remains unresolved, some infringement claims related to the output of generative AI systems have been dismissed due to insufficient evidence that the content generated by AI resembled copyrighted works.
