This is incredibly misleading. OpenAI internal systems were pwned, and in all cases, the labs absolutely are responsible for their models.

Yes, vendors are also irresponsible, but this misses the point.