Would the automatics eventually recover from any condition resulting from a jet upset? Stall/Spin? Inverted?
Because I suspect this will come up over and over I am thinking about putting a write-up on tech log about automation challenges and issues, and the direction I think things are likely to go.
I think we should remember that automation in aircraft is still really in its infancy and we really haven't figured out how best to integrate pilots into the system. There are a very large number of issues that need to be approached.
Although I tend to be fairly critical of Airbus for a few details (lack of feedback of sidestick inputs is a big one), I think the overall approach of flight laws is likely to be expanded over time in the name of safety, and that really solving the issues of automation are likely to mean better encapsulation and separation of actual duties. This means both giving more tasks back to the pilots (to keep them engaged and in the loop) but also making the plane a bit more autonomous as well. This means better avoidance and recovery regarding upsets, stalls, untrusted instruments, etc.
This of course doesn't mean that you shouldn't have a way to kill a lot of the automation and fly the plane pretty manually (this may happen anyway in even the most sophisticated fighter aircraft -- see the F22/international date line incident).
Obviously there are aerodynamic limits. There are also informatic limits. There are physical limits. A lot of the value is likely to be in avoiding these limits since not much can be done to overcome them.