HarryMann
I didn't make myself clear. I understand the point about automatic switch off. I'm suggesting that the crew were not responsive. First of all, they did not take evasive action before entering the weather system. Second, if your suggestion about turbulence is correct, surely the crew would have responded and the autopilot would have manually disconnected.
The turbulence theory could persuade me if some aerodynamic effect triggered the pitot errors that caused (?) the automatic disconnect. But isn't there some delay in the automated decision making, during which the crew would be taking action leading to manual disconnect?