A Federal Judge Ordered OpenAI to Stop Deleting Data. Here's How That Could Impact Users' Privacy

A federal judge has ordered OpenAI to stop deleting output data from its popular large language model, ChatGPT. It's part of an ongoing lawsuit from The New York Times alleging that OpenAI violated copyright laws by using "millions" of articles published by the newspaper to train its AI model, which now directly competes with the Times' content as a result. OpenAI is required to hold on to data The judge's order intends to prevent OpenAI from intentionally deleting information that could provide relevant evidence to the Times' allegations. The move could also have unintended consequences related to user privacy, experts said. The order effectively forbids OpenAI from deleting output data of chat logs, noted Jacob Flowers, data and privacy lawyer at Campbell Teague. That means that even if a user is deleting their own conversations with ChatGPT or using the chatbot on temporary mode, the logs could still end up in the hands of the Times' lawyers. OpenAI filed a motion to reconsider the order on May 15, arguing that keeping all of its output logs is way more information than is necessary, and that most of it wouldn't be relevant to the case anyway. The judge denied that motion, however, and directed the company to meet with the Times on May 27 in front of a magistrate judge to discuss what level of data retention would make sense. "The fact that the court took time to appreciate the technological complexities necessary to preserve this data ... is worrisome," Flowers told ADWEEK. "Those contractual obligations include the obligation to respect users' privacy, which the court appears [to have] let fall by the wayside without a second thought rather than impose any kind of duty on the plaintiffs to show the retention of this data would advance the litigation in a meaningful way." OpenAI declined to comment for this story, directing ADWEEK to its reconsideration request letter. The judge's order doesn't violate privacy laws like the California Consumer Privacy Act and Europe's General Data Protection Regulation, both of which have carveouts for legal requirements like this lawsuit. However, the action could have other implications, noted Daniel Goldberg, a data lawyer at Frankfurt Kurnit Klein & Selz. As a result of the order, OpenAI is "required to hold a lot more data than normally they would be," Goldberg told ADWEEK. That could make OpenAI more susceptible to security breaches, or shake the trust of consumers who expected their chatbot records to be deleted. There are also potential implications regarding energy use, storage, and environmental impact that the judge may not have considered when making the order, Goldberg said. "There will be people, as a result of this, that are concerned about what this means for working with these large technology providers," he said. The lawsuit is happening as tech majors like Meta, Amazon, and Google appear to be backtracking on data privacy initiatives. Meta has picked its facial recognition work back up, Google scrapped its plans for a cookie-free privacy sandbox, and Amazon stopped letting Alexa users opt out of having their data sent back to the company to train AI models, according to reporting from The Information.

AI News

A Federal Judge Ordered OpenAI to Stop Deleting Data. Here's How That Could Impact Users' Privacy

The New York Times is suing OpenAI for alleged copyright infringement

OpenAI is required to hold on to data

Kathryn Lundstrom

About

Subscriptions

Events

Publications