OK let me concede (for the moment
) that evolution has bestowed the most perfect means of sensing the exact direction of the magnetic lines of force passing through the bird's brain. The part I'm having touble with is that except for flight in the vicinty of the equator, if the bird banks far enough there is no useful information to sense. IOW if you had an electronic version of the bird's magnetic sensor(s) you couldn't keep an airplane right side up in a moderately steep bank without something else (like a gyro) to provide some indication of your attitude, nor could a supercomputer driven autopilot.
Similar problem with "gravity". It's not possible to distinguish between acceleration due to mutually attracting masses and that due to an external force (like lift) applied to one mass. Doesn't matter if your "gravity sensor" is organic or not, there isn't anything that can be sensed to distinguished.
Now one thing I can believe is that some birds may be able to revert to a stable falling body state much like a person hanging from a parachute if they "lose it" in IMC and then pull out before hitting the ground or even descend at a sufficiently low velocity that they survive the impact with the ground if they can't pull out.
If you can't buy this, we'll just have to agree to disagree till some facts come along.