OpenAI Resists Order to Release 20 Million Private ChatGPT Chats
▼ Summary
– OpenAI is appealing a court order requiring it to provide 20 million user chat logs to The New York Times and other plaintiffs in a copyright lawsuit.
– The company argues these complete conversation logs would expose significant private user information, unlike individual prompt-output pairs.
– OpenAI states over 99.99% of the requested chats are irrelevant to the case and has proposed alternative methods for identifying relevant logs.
– The New York Times is seeking the chats to find examples of users bypassing its paywall using ChatGPT, according to OpenAI’s public statement.
– Beyond this case, ChatGPT conversations have appeared in Google search results, prompting OpenAI to announce plans for enhanced security features including client-side encryption.
OpenAI is pushing back against a judicial mandate to release a massive collection of private user conversations from ChatGPT, a move initiated by The New York Times and other media entities pursuing a copyright lawsuit. The artificial intelligence firm argues that the court’s directive to hand over 20 million user chats is excessively broad and poses a significant threat to individual privacy. While OpenAI had previously suggested providing this smaller dataset as a compromise, far less than the 120 million records initially demanded, it now contends that the ordered disclosure goes too far.
In legal documents submitted to the U.S. District Court for the Southern District of New York, OpenAI emphasized that these logs capture entire conversational threads between users and ChatGPT. Each record includes multiple prompts and the AI’s corresponding responses, forming a comprehensive narrative of the exchange. The company stated that releasing such detailed records would be comparable to eavesdropping on a full discussion rather than hearing just a brief snippet, dramatically increasing the risk of exposing sensitive or confidential user information.
OpenAI’s court filing further noted that the overwhelming majority of these conversations, more than 99.99%, are completely unrelated to the copyright allegations at the heart of the lawsuit. The company has formally requested that the district court withdraw its current order and instead require the news organizations to engage with OpenAI’s own proposal for identifying and producing only those logs that are directly relevant to the case. Should the district court deny this request, OpenAI retains the option to escalate the matter to a federal appeals court for further review.
In a public message directed at its user community, OpenAI clarified that The New York Times is seeking access to these private conversations primarily to uncover instances where individuals may have used ChatGPT to bypass the newspaper’s digital paywall. This disclosure has heightened existing concerns among ChatGPT users regarding the privacy and security of their interactions with the AI assistant.
Privacy issues extend beyond this specific litigation. For example, some ChatGPT conversations have previously appeared within Google search results and within the Google Search Console, a tool available to developers for monitoring website traffic. In response to these broader concerns, OpenAI has announced its intention to introduce more sophisticated security measures. These planned enhancements include implementing client-side encryption for all messages exchanged with ChatGPT, aiming to provide users with stronger assurances that their data will remain confidential and protected from unintended exposure.
(Source: Ars Technica)
