OpenAI is now required to store ChatGPT user data, including deleted conversations and API data, due to a court order tied to the ongoing copyright lawsuit brought by The New York Times.
The New York Times is demanding that OpenAI permanently retain all ChatGPT and API user content, even if users have already deleted it. The publication says this is necessary to preserve potential evidence for the lawsuit.
OpenAI has pushed back against the demand, calling it excessive and at odds with established data protection standards. COO Brad Lightcap described the order as an "overreach" that could undermine user trust. Normally, OpenAI deletes user content after 30 days.
Who is affected?
The order impacts users of ChatGPT Free, Plus, Pro, and Team, as well as API customers who do not have a zero data retention agreement. Users of ChatGPT Enterprise, ChatGPT Edu, and API customers with zero data retention enabled are not affected.
OpenAI says that the data covered by the order is stored on secure systems under a legal hold, with access limited to a small, vetted group in the legal and security departments, and only for compliance purposes. Automatic disclosure of user data to the plaintiffs is not permitted.
The company has filed a motion for review by the magistrate judge and appealed the order to the district court, but will comply with the legal requirements until a decision is made.
According to OpenAI, the order does not change existing training policies. Data from business customers will not be used for model training, and individual users can still choose whether their content is used to improve OpenAI's models.
