Comment by otterley
5 hours ago
This article says nothing of the sort. The court order is to preserve existing logs they already have, not to disable logging, and hand all the logs over the plaintiffs. OpenAI's objections are mainly that 1/there are too many logs (so they're proposing a sample instead) and that 2/there's identifying data in the logs and so they are being "forced" to anonymize the logs at their expense (even though it's what they want as a condition of transferring the logs).
There is nothing in the article that mentions OpenAI being forced to create new logs they don't already have.
[flagged]
If OpenAI truly didn't keep conversation records for any length of time, they would not be subject to this kind of order. Lots of stateless services get these and are able to defeat them because they never store the user's data. The fact that they store them at all means that they are in scope for a preservation order. It also means that they are in scope for all manner of usage by OpenAI themselves even if a user requests deletion.
It seems as if the court has forced OpenAI into collecting logs that they weren't otherwise collecting, or that they were deleting at user request.
So in this case not keeping logs as ordered by the court would be contempt of court.
1 reply →