Federal magistrate judge directs OpenAI to preserve output log data regardless of users' requests or privacy laws and regulations

Issue 30 

On Tuesday, May 13, 2025, in the ongoing AI-related copyright infringement litigation against OpenAI, the Hon. Ona T. Wang issued an order directing OpenAI “to preserve and segregate all output log data that would otherwise be deleted on a going forward basis until further order of the Court (in essence, the output log data that OpenAI has been destroying), whether such data might be deleted at a user’s request or because of ‘numerous privacy laws and regulations’ that might require OpenAI to do so.”[i]  

On the same date, the Court scheduled a Status Conference on May 27, 2025 regarding deletion of output log data.[ii]  

Background of the Dispute 

OpenAI’s deletion of certain output log data was first raised with the Court by the New York Times consolidated case group (consisting of The New York Times Company, Daily News, LP et al., and The Center for Investigative Reporting, Inc. cases) plaintiffs (hereafter the “News Plaintiffs”) in January 2025.[iii] At that time, the Court denied the News Plaintiffs’ request for comprehensive preservation of OpenAI’s output log data.[iv]  

On March 25, 2025, the Daily News and The New York Times filed a letter motion to compel OpenAI to identify the nature and extent of output log data destroyed since the New York Times filed its complaint.[v] OpenAI filed a response on March 31, 2025, indicating that it had provided details of the tens of billions of ChatGPT Free, Pro, and Plus user conversations that had been retained, as well as the estimated fraction of those user conversations that had not been retained.[vi] Further, OpenAI reported that it had confirmed that the ChatGPT Free, Pro, and Plus conversation data that had not been retained was not retrievable.[vii] Additionally, OpenAI explained that it does not track the number of ChatGPT Enterprise conversations or application programming interface (“API”) completions that have not been retained, and that such data was not retrievable.[viii]  

On April 3, 2025, the United States Judicial Panel on Multidistrict Litigation issued a transfer order to centralize twelve actions against OpenAI defendants in the United States District Court for the Southern District of New York before the Hon. Sidney H. Stein, which included the New York Times consolidated case group.[ix] On April 30, 2025, the Hon. Sidney H. Stein issued Case Management Order No. 1 for the centralized group of twelve actions against OpenAI, directing, among other things, that “each party shall take reasonable steps to preserve all evidence that may be relevant to the ligation.”[x]  

On May 7, 2025, the News Plaintiffs sought leave from the Hon. Sidney H. Stein to file a supplemental letter to the Magistrate Judge Ona T. Wang regarding their pending motion to compel OpenAI to identify output logs that had been destroyed since the New York Times filed its complaint, in light of the Court’s Case Management Order No. 1 of April 30, 2025.[xi] OpenAI opposed the News Plaintiffs’ request to file the supplemental letter.[xii] The Hon. Sidney H. Stein granted the request to file the supplemental letter on May 9, 2025.[xiii]  

On May 12, 2025, the News Plaintiffs filed the supplemental letter to the Hon. Ona T. Wang, requesting the rescheduling of a discovery hearing on its March 25, 2025 letter motion to compel.[xiv] Portions of the letter were redacted, however, the News Plaintiffs argued that the volume of ChatGPT Free, Pro, and Plus conversations that have been destroyed was “substantial,” as well as that OpenAI has refused to provide the volume of output log data retained or destroyed for other OpenAI products at issue in the litigation, including the ChatGPT Enterprise and API products.[xv] The letter also indicated that in response to Case Management Order No. 1, the News Plaintiffs had renewed their request to OpenAI to stop destroying potentially relevant output log data.[xvi]  

The following day, the Hon. Ona T. Wang issued the orders described above.[xvii]  

OpenAI’s Request for Reconsideration 

On May 15, 2025, OpenAI filed a request for reconsideration of the Court’s order.[xviii] While portions of the request for reconsideration are also redacted, OpenAI represented that it was taking steps to comply with the Order, but explained that complying with the order raised many practical challenges, while requiring OpenAI to disregard its commitments to hundreds of millions of people, businesses, governments, and educational institutions.[xix] Further, OpenAI argued that there was no basis for the Court to find that the data covered by the Order was relevant, and that the News Plaintiffs’ letter contained factual misstatements that 1) OpenAI had deleted data in response to litigation, and 2) that the News Plaintiffs had only recently learned of their data retention policies.[xx] OpenAI argued further that while the News Plaintiffs have characterized OpenAI as destroying data, in reality, OpenAI does not retain data that 1) its users do not elect to retain, or 2) it has agreed not to retain under negotiated agreements or general terms with its customers.[xxi] OpenAI requested that the Court vacate the Order or stay the order pending the hearing on May 27^th.[xxii] 

What the Court Order Means For Lawyers Directly or Indirectly Utilizing OpenAI’s Products 

Lawyers using OpenAI’s products (such as ChatGPT) in their practices may wish to review the Court’s order directing OpenAI to preserve and segregate all output log data and consider whether to modify their organization’s AI usage policies. Additionally, lawyers using other AI tools in their practices may wish to confirm whether any of those AI tools utilize an OpenAI model via an API, and if so, they may wish to reach out to their AI tool vendor(s) for clarification 1) of the vendor’s data agreements with Open AI, and 2) what the Court’s order means for their data, so they can determine whether to modify their AI usage policies in light of the Court’s order.

I’ll share future developments on LinkedIn and in this newsletter.  

Thanks for being here. 

Jennifer Ballard 

Good Journey Consulting

P.S. Please consider sharing this issue of the newsletter if you know a lawyer who could benefit from reading it. Thank you!

[i] Order at 2, In Re: OpenAI, Inc., Copyright Infringement Litigation., No. 25-md-3143 (S.D.N.Y. centralized Apr. 3, 2025), document in relation to The New York Times Company v. Microsoft Corporation et al., No. 1:23-cv-11195 (S.D.N.Y. filed December 27, 2023). 

[ii] Amended Order at 1, In Re: OpenAI, Inc., Copyright Infringement Litigation. 

[iii] Order at 1, In Re: OpenAI, Inc., Copyright Infringement Litigation. 

[iv] Id. 

[v] Letter Motion to Compel at 1, The New York Times Company v. Microsoft Corporation et al., No. 1:23-cv-11195 (S.D.N.Y. filed December 27, 2023). 

[vi] Letter Response to Motion at 1, The New York Times Company. 

[vii] Id. 

[viii] Id. 

[ix] Transfer Order at 1, In Re: OpenAI, Inc., Copyright Infringement Litigation. 

[x] Case Management Order No. 1 at 4, In Re: OpenAI, Inc., Copyright Infringement Litigation. 

[xi] Request for Permission to File Supplemental Letter Regarding News Plaintiffs’ Motion to Compel OpenAI to Identify the Output Logs It Has Destroyed at 1, In Re: OpenAI, Inc., Copyright Infringement Litigation. 

[xii] Letter to the Hon. Sidney H. Stein dated May 8, 2025, In Re: OpenAI, Inc., Copyright Infringement Litigation. 

[xiii] Memo Endorsement Granting Request for Permission to File Supplemental Letter at 2, In Re: OpenAI, Inc., Copyright Infringement Litigation. 

[xiv] Letter Reply to Response to Motion at 1, In Re: OpenAI, Inc., Copyright Infringement Litigation. 

[xv] Id. at 1, 3. 

[xvi] Id. at 1, n.1. 

[xvii] Amended Order, Order, In Re: OpenAI, Inc., Copyright Infringement Litigation. 

[xviii] Reconsideration of Order Directing OpenAI to Preserve Output Logs, In Re: OpenAI, Inc., Copyright Infringement Litigation. 

[xix] Id. at 1, In Re: OpenAI, Inc., Copyright Infringement Litigation. 

[xx] Id. 

[xxi] Id. at 3. 

[xxii]Id. at 4.