OpenAI: NYT ‘Hacked’ ChatGPT for Copyright Case
OpenAI has approached the federal court in Manhattan requesting the dismissal of certain claims by The New York Times, accusing the newspaper of “hacking” its ChatGPT chatbot and other artificial intelligence systems to fabricate misleading evidence for the case.
It should be noted that the lawsuit against OpenAI and its largest financial partner, Microsoft, filed at the end of December, alleges that the companies utilized millions of New York Times (NYT) articles without permission to train their language models. These models now operate with facts from these articles and compete with news agencies as a source of credible information.
This case is not the first instance where copyright owners have accused technology companies of unlawfully using their materials for AI training. For example, Getty Images previously sued Stability AI for the unauthorized use of its images to train a generative neural network.
Regarding ChatGPT, OpenAI claims that the NYT coerced the neural network into reproducing its materials using “deceptive queries that explicitly violate OpenAI’s terms of use.”
In its statement, OpenAI criticizes the newspaper for claims that do not match their “renowned standards of journalism” and accuses it of hiring an external party to manipulate AI systems, though it does not directly accuse the newspaper of breaking laws.
In response, NYT’s lawyer Ian Crosby stated that actions described by OpenAI as “hacking” are legal uses of the company’s products to seek evidence of copying and reproduction of the newspaper’s copyrighted materials.
Technology companies often assert that their AI systems fairly use copyrighted material or do not use it at all. However, it is generally impossible for third parties to verify this accurately.
Undoubtedly, such lawsuits, while intended to protect copyright owners, also threaten the growth of the potentially multi-billion-dollar AI industry, which represents the digital future.
The question of whether training AI on copyrighted materials qualifies as “fair use” has not yet been resolved by the courts. Meanwhile, OpenAI, in its statement, expressed confidence in victory based on the question of fair use.
OpenAI emphasized that the newspaper cannot prevent artificial intelligence from acquiring knowledge about facts, just as any other news organization cannot forbid the NYT from reprinting stories in which it did not directly participate.
