First off I have to say that personally, I believe there have been more than 2 documented cases of the ILS autolock phenomenon. Is it rampant? No I will grant you that but I do not discount the number of cases you do based on the assumption that people can not adequately diagnose their gun's failure or have an ax to grind. To be frank the truth probably lies somewhere in the middle, however your blanket discounting of other reports makes it seem that you sir are the one with the ax to grind.
Second any real estimate of the frequency of autolock would have to be based on actual use of the gun, not production numbers. How many guns have the lock disabled? How many end up loaded and tossed in a night stand with no rounds fired. As I said prior no one knows, not you, not I and I admit that the number is small. You make sweeping generalizations and assumptions and generate a small number, if I wanted to I could make my own sweeping generalizations and assumptions and come up with a number that is small as well but perhaps several orders of magnatude larger than yours. It's called playing with statisicis and it's not going to generate meaningful numbers. I admit the number is small to truly know how small requires information neither of us have.
Third my point about medicine and aerospace engineering is not about a specific device, my point is about how we make choices in these fields with things that need to not fail or else there are grave consequences. In this arena if we can eliminate something that happens some small percentage of the time or make that percentage even smaller this is considered a very real benefit. I, and many others feel the same about our guns, if we can eliminate this small number form our guns why would we not? And eliminating it does not require some expensive metallurgy or exotic engineering, just eliminate something that is not needed or used. Or as others have pointed out design it so failure is less likely, or so that it fails - safe.