Two AoA sensors is not the answer, which one is right? Airbus has three and already has one fatal crash where two AoAs were were iced up, and the computer decided they were right and the remaining unfrozen sensor was wrong.
Any sensors in the airstream are exposed to damage, ice etc, no guarantee that a flock of birds should take out just one AoA sensor.
Surely the software should track pitch output from the attitude gyros against AoA, and report any sudden deviation.