OpenAI has formally requested a Manhattan federal court to eliminate certain claims from a copyright infringement lawsuit brought by The New York Times. OpenAI alleges that the newspaper engaged someone to infiltrate its ChatGPT platform, resulting in the production of a hundred pieces of deceptive evidence. OpenAI contends that these actions, if true, are in breach of their product’s terms of service and should not be admissible in court.
Allegations of Manipulation and Misconduct
OpenAI’s legal response stops short of directly accusing The New York Times of illegal hacking but emphasizes the lack of adherence to journalistic standards in the newspaper’s complaint. They suggest that the truth will reveal the Times’ involvement in unauthorized access to their systems. On the other hand, Ian Crosby, the counsel for the newspaper, claims that the allegations are an attempt to unearth evidence to support their case that OpenAI unlawfully replicated copyrighted works.
The original lawsuit by The Times, initiated in December 2023, alleges that OpenAI, with support from Microsoft, used its vast collection of articles to train chatbots without permission. The Times invokes the U.S. Constitution and the Copyright Act to protect its journalism, specifically pointing out instances where Microsoft’s Bing AI replicated its content verbatim.
Broader Implications for AI and Copyright
The New York Times is not alone in its legal battle against tech companies over AI training methods. Other content creators, including authors, artists, and music publishers, have brought forth similar claims. OpenAI has previously stated that the inclusion of copyrighted content is essential for the development of sophisticated AI models, arguing this to be a necessity rather than a choice.
In discussions with the UK’s House of Lords, OpenAI explained the challenges of AI development within the confines of strict copyright laws and noted the potential threat these legal challenges pose to the burgeoning AI industry, which they claim could be worth trillions.