Internet rumour mill:
> > As soon as they did that, the computers automaticlly 'released' all the
> > brakes. ('this is a Safety feature so that pilots don't land with the
> > brakes on'.)
Ridiculous on the face of it - landing is precisely when a pilot *does* need the wheelbrakes.
Sounds to me like some bright spark has decided to cherry-pick the debate about the Hamburg and Conghonas incidents and tried to apply it to a completely unrelated scenario. Added to this the language has all the hallmarks of the most poorly-informed Airbus bashing ("fooling" the computer, the computer "thought" it was in the air or taking off). It doesn't even differentiate between spoiler braking (which is affected by the ground sensor, but is not relevant in this case) and wheel braking (which to the best of my knowledge is either not, or only marginally affected by the ground sensor and is very relevant in this case).
Finally - if you're going to pop a CB to shut the alarm off (unwise, but for argument's sake plausible), why would you waste time second-guessing the computer logic and pop the ground sensor when you could just pop the alarm? I'm committing a bugbear of mine by making an assertion of what a pilot would or wouldn't do there, but it just doesn't seem to follow to me.
The closest wild guess I would make is that the big hole in the cheese was human error of some kind, whether that error was in the cockpit (CRM's just as important in a test environment as in line flying), or a communication error on the ground.