Skip to content
IFAI

Industry

OpenAI fights court order to save all ChatGPT logs

Published Reach 44

OpenAI is fighting a court order requiring it to preserve all ChatGPT logs, including deleted chats and sensitive API data. The mandate followed accusations by The New York Times and other publishers that OpenAI was destroying copyright evidence. OpenAI called the sweeping order premature and warned it puts hundreds of millions of users' privacy at risk.

OpenAI fights court order to save all ChatGPT logs
Image: arstechnica.com

Every time you hit "delete" on a ChatGPT conversation, you assume it's gone for good. A federal judge just ruled otherwise.

OpenAI is currently fighting a controversial court order requiring it to preserve all ChatGPT user logs. That includes conversations users explicitly deleted, as well as sensitive queries sent through its commercial API offering.

In a fiery court filing, OpenAI requested oral arguments to block the order, calling the mandate an unprecedented threat to the privacy of hundreds of millions of users worldwide.

Save everything, even the deleted chats

The whole mess traces back to a May 13 order issued by Magistrate Judge Ona Wang.

The court directed OpenAI to preserve and segregate all output log data that would otherwise be deleted on an ongoing basis. In plain English, the system is now legally barred from actually wiping out chat records.

The scope is massive. OpenAI confirmed that the preservation order hits ChatGPT Free, Plus, and Pro accounts, along with business customers relying on its developer API. On Thursday, the company even rolled out an FAQ to explain to nervous users who is affected and how their data could potentially be accessed.

Copyright cops want your browser history

Why did the court step in? Blame the ongoing copyright fight spearheaded by The New York Times and other media plaintiffs.

The news organizations accused OpenAI of destroying evidence. Their theory is that people using ChatGPT to bypass news paywalls would naturally delete all their searches to cover their tracks. Because OpenAI previously only handed over sample logs from users who opted in to data retention, the publishers argued that crucial evidence was vanishing every single day.

Judge Wang bought that argument, concluding that OpenAI would likely keep deleting data unless forced by the court.

OpenAI fired back, arguing that the judge rushed the decision based on nothing more than a hunch. The company contends that stripping away user privacy controls without any proven need is wildly disproportionate, and the May 13 order should be vacated immediately.

Why it matters

This dispute marks a dangerous escalation where a corporate copyright spat directly collides with consumer and enterprise privacy.

For regular users and developers, it's a harsh reminder of how the cloud actually works. Hitting delete doesn't guarantee your prompt is gone if a court steps in. If your company feeds proprietary source code or sensitive customer data into OpenAI's API, having those interaction logs quarantined indefinitely by court order is an absolute compliance nightmare.

For the AI industry, the precedent is alarming. Forcing an AI vendor to preserve private user activity just to see if someone might have skirted a paywall looks like a massive fishing expedition. OpenAI is right to push back aggressively here; if this order stands, no privacy toggle in any AI product can ever truly be trusted again.

Sources (1)

  1. [1]
    arstechnica.com(opens in a new tab)Originalarstechnica.com

Best of AI, daily