OpenAI Was Breached in 2023. The Company Is Still Determining What Was Taken.

OpenAI experienced a breach in 2023 that remained undisclosed until later, when the company described it as limited without clarifying what limited means. The systems breached contained information about AI systems themselves. The responsible department has been identified internally but not named publicly.

Companies breaching their own infrastructure and delaying disclosure while minimizing scope is standard procedure now. The pattern is: incident, delay, redescription, partial transparency. No one files the obvious report that this is how it works.

The investigation into what was taken will continue indefinitely. Limited means the breach was large enough to acknowledge but small enough to contain narratively. By the time the scope becomes clear, the news cycle will have moved elsewhere and the department will have moved on.