PJ2, nice you're back.
I've noted the PF's reaction to the AP/AT drop out. As written in the writeup:
From 2 h 10 min 05, the autopilot then auto-thrust disengaged and the PF said "I have the controls". The airplane began to roll to the right and the PF made a left nose-up input. The stall warning sounded twice in a row. The recorded parameters show a sharp fall from about 275 kt to 60 kt in the speed displayed on the left primary flight display (PFD), then a few moments later in the speed displayed on the integrated standby instrument system (ISIS).
The $64,000 question that moment in time poses is whether that all happened at once or smeared, in the order stated, over 11 seconds. If it was smeared over time in the order written, I'd presume that would be uncertain airspeed. If it isn't what would be the PF reaction taken as he says he has the controls?
If they all happened in a slightly different time sequence the PF might have a whole different chain of actions initiated.
Another question is how good the simulator training is for cascade failures, which certainly appears to be what happened here. PF starts uncertain airspeed knowing he's going 275kts CAS. And a stall warning hits, a sure disruption to his chain of thought. Is training good enough that there is a smooth transition to realizing it's false and the pitots are out or does it start a whole new chain of trained reactions or does it leave the pilot rattled?
(I'm rather fastened on that spurious stall warning that seems to have happened more than just this once and left pilots transfixed trying to solve the impossible. This seems like one obvious change required in the flight control systems. If kCAS indication drops dramatically and that is not matched by a similar reaction in the inertial reference unit, suppress the stall warnings even if they might be real as might be the case with 10 degree AOA and 268 kts. It a pilot knows he won't get spurious stall warnings (or its far less likely) then a real stall warning will be heeded properly. I'm simply relying on my computer training here. You can overwhelm systems with conflicting input. In fact it's a known hacking technique. I also think this problem may well have already been dealt with in the flight controls for the new fully autonomous drones.)