OpenAI has requested a federal judge to dismiss certain portions of the New York Times’ copyright lawsuit against the company, asserting that the newspaper engaged in “hacking” of its ChatGPT chatbot and other artificial intelligence systems to produce misleading evidence for the case.
In a filing submitted to the Manhattan federal court on Monday, OpenAI claimed that the Times induced the technology to replicate its content through “deceptive prompts” that clearly violated OpenAI’s terms of use.
OpenAI argued that the Times’ allegations did not adhere to the newspaper’s renowned journalistic standards and contended that the truth, which would be revealed during the case, is that the Times paid someone to hack OpenAI’s products.
As of now, representatives for both the New York Times and OpenAI have not responded to requests for comments on the recent filing.
The lawsuit against OpenAI and its major financial supporter Microsoft, filed by the Times in December, accuses them of utilizing millions of articles without authorization to train chatbots for providing information to users.
The New York Times is one of several copyright owners pursuing legal action against tech companies for the alleged improper use of their content in AI training. This includes groups representing authors, visual artists, and music publishers.
Tech companies have consistently argued that their AI systems make fair use of copyrighted material and have emphasized that these lawsuits pose a threat to the growth of a potential multitrillion-dollar industry.
The complaint against OpenAI and Microsoft accuses them of attempting to benefit from the Times’ substantial investment in journalism and creating a substitute for the newspaper. It points to instances where OpenAI and Microsoft chatbots provided users with nearly identical excerpts from Times articles when prompted.
In response, OpenAI, in its filing, stated that it took the Times “tens of thousands of attempts to generate the highly anomalous results” and emphasized that, under normal circumstances, ChatGPT cannot be used to deliver Times articles at will.
