> AND the user did not intend to have the model act in an illegal matter.

I find this an assumption that is not based on any facts. The user did not provide instructions to follow nor to break laws, so if you look at it from a computer (that does not make assumptions) standpoint, there is no rule to follow there thus it can do what will create the best possible outcome for the task.

We presume innocence where I'm from. If a message contains nothing nefarious, nefarious intent can not be assumed.