I’ve seen both [maybe you think random implies someone just didn’t fess up] types of errors. Certainly human errors are the vast majority but there are plenty of examples otherwise where no human had an opportunity to intervene.
The question is not about what kind of errors you've seen, but how they have been initially described.
One more time: Have you ever seen errors described as (and at the time appear to be) "random", and then turn out not to actually be random when more information is available?