OpenAI is now seeking to dismiss the copyright lawsuit from The New York Times by insisting that the publisher hired someone to hack into ChatGPT to obtain evidence against the AI firm.
The AI startup wrote on its court filing on Monday that the Times "thousands of attempts to generate anomalous results" by "targeting and exploiting a bug" within the chatbot's system.
OpenAI accused the publisher of using deceptive methods that "blatantly violate OpenAI's terms of use." The company insisted that no "normal people" use its AI tool as the Times did.
The Times Deny 'Hacking' Allegations from OpenAI
According to CNBC News, the "hacking" OpenAI claimed is called prompt engineering or "red teaming," a stress testing technique used by safety teams and academics to look for vulnerabilities in AI.
It is worth noting that the company did not accuse The Times of breaking any anti-hacking laws.
The Times dismissed OpenAI's allegations in a statement on Tuesday, saying that OpenAI "bizarrely mischaracterizes" the definition of hacking and that The Times is "simply using" the company's technology to look for copyrighted works.
The newspaper company filed a lawsuit against OpenAI last December for supposedly using its licensed works to
OpenAI Seeks to Junk The New York Times Lawsuit
This was the second time OpenAI has questioned the legitimacy of the evidence the newspaper company provided for the accusation of its content being used to train ChatGPT.
Last month, the company accused The Times of manipulating prompts of ChatGPT to show that the chatbot is giving verbatim results of Times articles.
The news publication also dismissed the allegations, maintaining its position that the results still prove that OpenAI has access to "the enormous aggregate of human knowledge" within the Times's catalog.
On the other hand, OpenAI CEO Sam Altman admitted that the company's web crawler might have taken data from The Times by accident and that it is opting out of using it throughout the lawsuit.
Several of the content The Times claims that OpenAI has stolen was locked behind paywalls, something that is not easily obtainable by bots alone.
Related Article : OpenAI Opts Out Using New York Times Data Amid Lawsuit, Altman Reveals
