If there is disagreement, it obviously doesn't matter which one is wrong to detect there is a malfunction. This case would warrant the disconnection of any automatic system that could rely on or even be influenced by those sensors, and issue a warning. First thing that happens is usually the disconnection of the autopilot, and I'm pretty sure that happens when there is a faulty pitot, at least on Airbus planes. This is the first thing that happened on the AF447 for instance, AFAIK. Of course, this could be different in this case. Maybe.
In the AF447 scenario, the computer decided it did not have reliable air data, so it not only disconnected the autopilot, it dropped into alternate law. Basically, the computer just punted and said "your airplane!" This might seem like reasonable behavior, but one might imagine that handing a degraded airplane to a pilot at the very worst moment is not a satisfactory failure mode, and in the AF447 scenario the outcome was obviously bad.
What else could the computer have been programmed to do? It could have perhaps dropped into an attitude-based straight-and-level flight mode, and made an annunciation inviting the pilots to take over, but not insisting that they do so that very instant.
But it's easy to second-guess these sorts of systems. It's not nearly as easy to design them so that they always do the thing that seems smartest in retrospect.
Any redundant sensor system that would give the same figures but all wrong, and that those wrong figures could still pass as valid would be severely badly designed IMO, or this would be a case of extreme bad luck.
I think you are underestimating the difficulty of this problem, because unusual attitudes and airspeeds are going to generate odd readings, and these are the exact situations -- not totally predictable -- where an AOA system is interesting.
Yes, one solution would be to have an array of AOA sensors from different vendors using different technologies, but you have to draw the line somewhere and you still need to integrate the data, and you still have to deal with the possibility that sensor #n, giving a strange reading is giving a correct strange reading. Maybe one sensor is at the wing root and another is at the tip, and the root is stalled and the tip is not. Or one wing is stalled and the other is not. Or both roots and one tip, but not the other, etc. Maybe the wing is physically damaged in some way that affects airflow and the sensor is correctly reporting that. All of those are aerodynamic situations that can happen.
Usually the different sensors should be located at different spots, ideally from different vendors, etc, so this should be rather unlikely.
But if they are in different spots, they're measuring different things. And "unlikely" is the exact problem. These aircraft are going to be flown 10's of thousands of hours and most of them are never going to ever get near a stall condition in their entire operational lives. Everything interesting an AOA system does is in well into the regime of "unlikely."