For example we do every ILS/GLS as dual channel approach, even if we plan to disconnect the autopilot later on for a normal manual landing
Now that's a new theory. I must say it seems like jumping at your own shadow. Surely the autopilot airworthiness people as well as the aircraft manufacturer would have considered the probability factor of an obscure defect such as an aileron channel runway and catered for that in the FCTM - if of course there was a problem in the first place.