A federal judge has terminated the controversial data preservation order that required OpenAI to indefinitely retain all ChatGPT conversation logs, marking a significant shift in the ongoing copyright litigation between the AI company and The New York Times.
Federal judge Ona T. Wang filed a new order on October 9 that frees OpenAI from its obligation to “preserve and segregate all output log data that would otherwise be deleted on a going forward basis."[5] This decision effectively reverses a May 2025 mandate that had sparked intense debate about user privacy and data retention in the artificial intelligence industry.
Background of the Legal Battle
The case originated in late 2023 when The New York Times filed a lawsuit against OpenAI for copyright infringement, alleging that the AI giant trained its language models using the news outlet’s intellectual property without proper authorization or compensation.[5] The litigation represents one of the most significant legal challenges facing generative AI companies as they navigate copyright law in the context of large language model training.
In May 2025, Magistrate Judge Wang issued a sweeping directive requiring OpenAI to retain and segregate all output log data, including conversations that users had explicitly deleted.[1] The plaintiffs argued that ChatGPT users could be reproducing paywalled content like news articles and then deleting those conversations to erase evidence of potential copyright violations.[2]
OpenAI’s Response and Appeal
OpenAI vigorously contested the preservation order, calling it an “overreach” that posed significant risks to user privacy.[5] The company argued that the decision was made hastily based on speculation rather than concrete evidence, and that it created what OpenAI characterized as a “privacy nightmare” conflicting with global data protection standards and user privacy expectations.[1]
The company also raised concerns about the technical and ethical implications of serving as a permanent data archive for millions of daily conversations. According to OpenAI’s legal filings, the mandate was both technically burdensome and potentially incompatible with regulations like the General Data Protection Regulation (GDPR).[1]
What Changed in the New Ruling
The October 9 order introduces important nuances to the data retention requirements. While OpenAI is now freed from preserving all future chat logs as of September 26, several key provisions remain in effect:
Preserved Historical Data: Any chat logs already saved under the previous May order remain accessible and must be retained.[5]
Flagged Accounts: OpenAI must continue to hold data related to specific ChatGPT accounts that have been flagged by The New York Times as potentially relevant to the copyright investigation.[5]
Expandable Scope: The New York Times retains the right to expand the number of flagged users as it continues examining OpenAI’s preserved records.[5]
Implications for ChatGPT Users
This development has significant implications for user privacy. During the period when the original preservation order was in effect—from May through September 2025—all ChatGPT conversations were retained indefinitely, even those that users believed they had permanently deleted.[1]
The order specifically affected users on ChatGPT’s consumer tiers, including Free, Plus, Pro, and Team subscriptions. However, it explicitly exempted ChatGPT Enterprise, ChatGPT Edu, and API customers operating under specific data governance frameworks.[1]
With the termination of the blanket preservation requirement, standard data retention policies can now resume for most users. However, anyone whose account has been flagged by The New York Times will continue to have their data preserved regardless of deletion attempts.
The Broader Context
This case highlights the complex intersection of artificial intelligence development, intellectual property rights, and user privacy in the digital age. As AI companies face increasing scrutiny over their training data sources, courts are grappling with how to balance discovery rights in litigation against individual privacy protections.
The outcome also sets a precedent for how future copyright disputes involving AI companies might be handled, particularly regarding what constitutes reasonable data preservation for legal proceedings versus undue burden on both the companies and their users.
For now, the modified order strikes a middle ground—allowing OpenAI to resume normal operations while preserving the plaintiffs’ ability to investigate specific accounts of interest. As the underlying copyright case continues to unfold, this data retention issue may resurface depending on what evidence emerges from the examination of flagged accounts.
Sources:
[1] https://magai.co/openai-court-ordered-data-retention-policy/
[2] https://www.youtube.com/watch?v=u3_vIBXrS0E
[6] https://openai.com/index/response-to-nyt-data-demands/