OpenAI alleges that The New York Times exploited a bug to manipulate ChatGPT and fabricate a lawsuit against the company.
About two months have passed since The New York Times filed a lawsuit against OpenAI, the creator of ChatGPT, and its backer, Microsoft. The lawsuit alleges that OpenAI and Microsoft “copied” articles, reports and other published content from The New York Times to train large language models (LLMs) that use the ChatGPT chatbot.
In a recent development in the legal process, OpenAI is now seeking to dismiss certain parts of the lawsuit. First reported by Reuters, OpenAI claims The New York Times “hacked” ChatGPT to obtain misleading evidence to support its claims. However, OpenAI has not provided information about the specific person or entity responsible for the alleged hacking of ChatGPT.
“They were only able to do so by targeting and exploiting the bug using deceptive prompts that blatantly violate OpenAI’s terms of service,” OpenAI said, asking for the case to be dismissed.
Additionally, OpenAI added that The New York Times defrauded ChatGPT by sending “fraudulent prompts that blatantly violate 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 during this case is that the Times paid someone to hack OpenAI’s products.”
The New York Times quickly responded, with its attorney Ian Crosby saying, “What OpenAI oddly characterizes as ‘hacking’ is simply using OpenAI’s products to look for evidence that they stole and copied The Times’s copyrighted work.”
To the uninitiated, the lawsuit against OpenAI is not limited to The Times, but that OpenAI “engaged in extensive copying” of many media organizations, but the NYT’s content received “special weight.”
Big tech companies in the AI sector have repeatedly stressed that AI systems use copyrighted material fairly, and that this lawsuit robs the multi-trillion dollar industry of potential growth. It is important to emphasize that currently the legislators have not provided clarification on whether AI training is classified as fair use of copyrighted material or not.
