Try this, "drivers are a threat to our road system." They clog it up and very often they crash into each other and cause serious issues to traffic. We need to protect the road system against *drivers*. Can we automate *cars* so they work without *drivers* as much as possible?"
Lo and behold, Google and any number of other entities are working on this very problem.
Except that that's not a valid analogy.
Automobile-based transportation systems (consisting of road, car, and car occupants) will, indeed, work just fine once we have made the cars run without drivers.
But if you remove the user from the equation of computer security, suddenly all you have is a bunch of perfectly secure computers that no longer have any purpose to their existence.
The reason we have computers is so that people can use them to perform a variety of tasks. It is fundamentally impossible to remove the user from the equation while still achieving the desired result—unless you have become so skewed with tunnel vision as to believe that the desired result is a perfectly secure computer.
The result we should all be aiming for is a computer that can perform the tasks required of it by its users without them running the risk of compromising security through their activities.