OpenAI, the research laboratory behind the popular language model ChatGPT, has filed a motion to dismiss parts of a copyright lawsuit filed by The New York Times. In a surprising twist, OpenAI claims that The New York Times “hacked” ChatGPT by employing “deceptive prompts that blatantly violate OpenAI’s terms of use” to generate evidence for the lawsuit.
The New York Times filed the lawsuit in December 2023, accusing OpenAI and its financial backer Microsoft of using millions of its articles without permission to train ChatGPT. This practice, known as text scraping, is a growing concern in the field of artificial intelligence, with copyright holders questioning the legality of using their content to train AI models.
OpenAI argues that The New York Times’s claims are misleading and that the evidence they presented was not gathered fairly. They allege that the Times used “malicious prompts and instructions” to manipulate ChatGPT into generating text that misrepresents the model’s capabilities and suggests copyright infringement.
The New York Times has not yet responded to OpenAI’s claims. This development adds a new layer of complexity to the ongoing debate about fair use and copyright in the age of AI. It raises questions about the ethical boundaries of testing and evaluating AI models, and the potential for manipulation when interpreting their outputs.
Further developments in this case are expected as both parties prepare for the upcoming court proceedings. The outcome could have significant implications for the future of AI development and the way copyrighted material is used in training AI models.
