OpenAI offers 20 million user chats in ChatGPT lawsuit. NYT wants 120 million.

OpenAI is gearing up to mount a significant defense against The New York Times, which is seeking access to a vast array of ChatGPT user logs. This case has drawn considerable attention as it could potentially reveal copyright violations within the AI's responses. In a joint letter submitted on Thursday, both parties expressed their desire to hold a confidential settlement conference on August 7. Legal representatives from the NYT confirmed that the conference's purpose is not to negotiate a settlement but to address one of the most contentious issues in the case: the examination of millions of ChatGPT logs by news plaintiffs. This development could soon clarify whether ChatGPT users' private conversations will be accessible during the lawsuit. In its ongoing efforts to restrict access, OpenAI has explained the intricate process required to make deleted chats retrievable, thereby contesting the NYT's push for wider access to the logs. The company has previously accused the NYT of attempting to conduct what it describes as "mass surveillance" over ChatGPT users. Despite their efforts, OpenAI faced a setback when it lost a bid to prevent news plaintiffs from accessing all ChatGPT logs. Following this defeat, OpenAI has shifted its strategy to minimize the number of logs available in the case while addressing growing privacy concerns among its users. For many, the lawsuit poses a risk of exposing sensitive information that OpenAI had previously assured would be deleted. In a recent proposal, OpenAI suggested a compromise to the court, arguing that news organizations should not be allowed to search through all ChatGPT logs. They referenced insights from computer science researcher Taylor Berg-Kirkpatrick, who indicated that a sample of 20 million logs could adequately determine the frequency with which users might be using the AI to bypass paywalls of news articles.