OpenAI has asked a federal judge to dismiss part of its copyright lawsuit against the New York Times, saying the newspaper used its chatbot ChatGPT and other artificial intelligence systems to generate misleading evidence in the lawsuit. claimed to have been “hacked”.
OpenAI said In a filing in Manhattan federal court. On Monday, the New York Times reported that the technology caused the material to be copied through “deceptive prompts in clear violation of OpenAI’s terms of service.”
“The allegations in the Times’ complaint do not meet its famously rigorous journalistic standards,” OpenAI said. “The truth that will emerge over the course of this litigation is that the Times paid someone to hack OpenAI’s products.”
OpenAI did not name the “hired guns” the Times used to manipulate its systems, nor did it accuse the Times of violating anti-hacking laws.
Representatives for The New York Times and OpenAI did not immediately respond to requests for comment on the filing.
In December, the newspaper accused OpenAI and its biggest financial backer, Microsoft, of using millions of the company’s articles without permission to train chatbots that provide information to users.
The Times is one of several copyright holders, including groups such as writers, visual artists and music publishers, who have sued tech companies for allegedly misusing copyrighted material in AI training.
Tech companies argue that their AI systems make fair use of copyrighted material and that the lawsuit threatens the growth of a potentially multitrillion-dollar industry.
Courts have not yet addressed the important question of whether AI training qualifies as fair use under copyright law. So far, judges have dismissed some infringement claims regarding the output of generative AI systems due to a lack of evidence that the AI-created content resembles copyrighted works. I have rejected it.
The New York Times’ complaint cites several instances in which OpenAI and Microsoft’s chatbots provided users with nearly verbatim excerpts of articles upon request. The newspaper accused OpenAI and Microsoft of “free riding on the Times’ huge investment in journalism” and trying to create an alternative newspaper.
OpenAI said in a filing to the Times that it took “tens of thousands of attempts to produce highly unusual results.”
“In the normal course of events, you cannot use ChatGPT to freely serve Times articles,” OpenAI said.
OpenAI’s filing also said it and other AI companies will ultimately prevail based on fair use issues.
“The Times cannot prevent an AI model from acquiring factual knowledge any more than it can prevent another news organization from re-reporting a story that the Times itself was not involved in investigating. ” said OpenAI.
