One thing does bother me about all this capacitive coupling. Wouldn't BOTH current carrying conductors couple to ground a similar amount and if so the net current differential across the CT in the GFCI would be close to zero? In fact the current in the ground wire would null too.
The current will be proportional to the voltage differential. For a 120V circuit, one conductor is at 120V while the other is at 0V with respect to ground. There will be current due to capacitive coupling from the hot conductor, but not from the grounded conductor, as it is at nearly the same voltage as the grounding conductor.
For a 240V circuit, the capacitive coupling would likely be balanced between the two current carrying conductors, as they are both at 120V.
