OpenAI no longer forced to save deleted chats—but some users still affected (arstechnica.com)

A US magistrate has approved a deal letting OpenAI stop preserving most deleted and temporary ChatGPT chats that had been retained under a court-ordered preservation requirement tied to a lawsuit by The New York Times and other news organizations. The original order forced OpenAI to save "all output log data that would otherwise be deleted" after plaintiffs alleged users were using ChatGPT to bypass paywalls; those saved logs captured ChatGPT’s outputs (not user inputs) and were reviewed by the news plaintiffs. OpenAI contested the mandate but initially lost; a joint motion filed by OpenAI and the news organizations was signed by Judge Ona Wang, allowing OpenAI to cease the practice as of September 26. Significance: the change eases a high-profile privacy and data-retention conflict at the intersection of AI service providers and civil discovery, setting a practical example of how preservation demands can be negotiated away even after litigation-driven retention. Technical implications include reduced long-term storage of model outputs that had been kept for potential evidentiary use, though the settlement leaves a carve-out—some users’ deleted or temporary chats will still be monitored or preserved—reflecting that targeted preservation for active litigation or specific investigatory needs can persist even when broad retention stops.