OpenAI, the artificial intelligence powerhouse behind ChatGPT, has announced that it is now required to indefinitely retain deleted user conversations due to a court order linked to an ongoing copyright lawsuit filed by The New York Times. The decision marks a significant shift in how user data is handled on the platform and has raised privacy concerns across the tech and legal communities.
A Court Order That Overrides Privacy Protocols
In a recent statement posted on OpenAI’s official blog, Chief Operating Officer Brad Lightcap shared that the company is being compelled to preserve “all output log data that would otherwise be deleted.” This includes chats that users have explicitly requested to delete, as well as data that may otherwise be subject to privacy-focused erasure under laws such as the GDPR or CCPA. The court has issued this directive as part of a discovery process intended to secure potential evidence for the New York Times in its legal battle against OpenAI and Microsoft.
Previously, OpenAI’s data retention policy stated that user-deleted chats would be stored for up to 30 days before being permanently deleted from its servers. That policy has now been suspended, pending further instruction from the court. This order affects all users of ChatGPT’s free, Pro, Plus, and Team versions. However, it explicitly does not apply to ChatGPT Enterprise and ChatGPT Edu customers, nor to clients with zero data retention agreements in place.
Legal Access, Not Public Exposure
OpenAI has emphasized that this temporary policy change does not mean deleted chats will become publicly accessible. Instead, the stored conversations will be locked down and accessible only by “a small, audited OpenAI legal and security team.” The purpose is strictly legal—to comply with the court order and assist in the litigation process, if necessary.
Still, the change raises alarms for privacy advocates and everyday users alike. The idea that conversations, once believed to be erasable, are now being stored indefinitely sits uncomfortably with many, especially in an era when digital privacy is already a major concern.
The Lawsuit: Copyright and Training Data
The legal dispute stems from a lawsuit filed by The New York Times in December 2023, alleging that OpenAI and Microsoft engaged in copyright infringement by using millions of the newspaper’s articles to train large language models like ChatGPT. The Times argues that the retention of user-generated data could serve as critical evidence, potentially showing that OpenAI’s models were trained on copyrighted content without permission or proper licensing.
While the lawsuit is ongoing, this move to preserve all user output data reflects how serious the discovery process has become. It also highlights the broader tension between intellectual property rights and AI training methods that rely heavily on vast swathes of publicly available online content.
OpenAI Pushes Back
OpenAI leadership has responded with strong opposition to the court’s mandate. CEO Sam Altman called the ruling “an inappropriate request that sets a bad precedent,” adding, “We will fight any demand that compromises our users’ privacy; this is a core principle.” OpenAI says it is actively appealing the court’s decision and will continue to push for user privacy protections.
Lightcap echoed similar concerns in his blog post, describing the court order as an “overreach” that “abandons long-standing privacy norms and weakens privacy protections.” Despite the legal pressure, OpenAI is attempting to maintain transparency with users while complying with the court’s requirements.
As of now, The New York Times has declined to comment publicly on the matter.
What This Means for Users
For the millions of ChatGPT users around the world, the implications of this court order are significant. It means that even if a user manually deletes a conversation, that data will still be preserved indefinitely—at least for the duration of the court’s directive. Although the data will not be made public and will be tightly controlled within OpenAI’s legal and security teams, the idea that “deleted” no longer truly means deleted has shaken trust among some users.
Users in regulated industries, or those dealing with sensitive or proprietary information, may need to reevaluate their use of ChatGPT’s non-Enterprise versions. Additionally, it underscores the importance of understanding what privacy protections are—or are not—in place when using AI tools, especially in a rapidly evolving legal environment.
A Precedent in the Making?
This situation represents a rare intersection of emerging AI technologies, legal accountability, and user privacy rights. Depending on how the courts ultimately rule, this could set a precedent for how AI companies handle data deletion requests in the face of legal scrutiny. It may also influence future legislation and regulatory efforts targeting AI data practices and intellectual property.
For now, OpenAI finds itself at a crossroads—balancing legal obligations, user expectations, and its own principles around privacy. As the case develops, users and industry stakeholders alike will be watching closely to see how the tension between transparency, compliance, and privacy plays out.
