I've seen metal on metal servo rods/arms cause interference on 27/29 Mhz radios from the vibration of the pushrod in the actuating arm, so a rotating shaft in a metal tube probably could cause the same issue, in saying that I have never grounded shafts in models.
It probably would be ideal practice though.
Full sized boats we have put a shaft brush connected to the rest of the bonding system, this reduces corrosion and destruction of bearings (through block to shaft stray currents).
I've often heard of this, but don't know of any physics that could cause it. Doing a search for RFI and metal to metal rubbing only comes up with model boat sites, so at a guess, something else is having an effect. The same search gives lots of RFI chat, but no mention of metal to metal.
To get into the radio via the aerial, you need something like the radio frequency or a harmonic, say 13MHz (million times per
second) for a 27MHz set, and enough power. A motor shaft is unlikely to rotate more than a few thousand times per
minute. The servo signals base on 50 per second, so that is a possible entry point, but I am still struggling to see how two bits of metal, disconnected from the wiring, rubbing against each other, are going to generate the kind of energy that will hop into the wiring with enough power to cause problems. A moving shaft rubbing insulation, yes, I can see that, but that is not apparently the problem. It would be nice to have a reasoned explanation rather than accepted folklore.