TestPilot, there are many documented cases of user error with non-safety-equipped devices causing bad outcomes, including fatalities. I am unaware of (and, at least a couple of years ago, Mas Ayoob stated he was also unaware of, IIRC) any instances where a bad outcome obtained because of a safety existing. The former seems to be a demonstrated risk. The latter seems theoretical. Perhaps that is because the former risk arises during times of relative relaxation and inattention, where as the latter ONLY happens when the focus is on making the gun go off.
First, I want to stress that I am not trying to convince people that no one should use guns with manual firing inhibitors. That is not my point.
If mistakes causing accident is possible, then mistakes causing manipulation failure of manual firing inhibitors are also possible.
There certainly are real life case of manipulation failures. The reason why you don't hear about them is that when accidents occur with Glock type guns, the result gets attention because a gun did fire, but when manipulation failure of a manual firing inhibitor occurrs, the result is gun not firing. In a training environment, where majority of it happens, it ends up nothing more than an embarassment, if it gets caught. On a range, the person who made the mistake will think "ooops" then flick the lever, then resume shooting. Other than an awkward split second delay, no one will even notice. However, a criminal who is trying to kill that person will notice.
Also, if we find a dead officer or deputy with a 1911, for example, in hand with the thumb lever in the engaged position, how is anyone going to determine whether if that officer or deputy failed to manipulate the lever or there was no attempt to manipulate the lever at all? Even if the gun was found in the dead person's hand with the lever down, how do we know if it was flicked down after a fumble that allowed the criminal to attack?
On the other hand, if a gun fired, whether if it shold or should not have, we will defenitely know. See the difference?
I've once landed on my palm when I fell during PT session in my military time, and when I had to practice with an M4 that afternoon it took at least a full second for me to flick the selector lever on an M4 because of that injury. Imagine being a second late in a close quartes pistol fight.
Instructor Rob Pincus even challenged people that if they go through his course with a sub-compact 1911 with no mechanical failures or manual firinig inhibitor manipulation mistakes he would refund the tuition. That will give you an insight into how often it actually happens.