This question has some implications on home related work with mains. Never take what you read on the internet as the truth when lives might be in the line. That is how I'll be treating potential answers also.
In an old household on some sockets the testing screwdriver lights up very a little when the ground is tested. The light is significantly weaker than on the hot wire.
The ground voltage can't be felt to the touch, the only way it's detectable is with the neon lamp screwdriver.
Other sockets don't have that leaking voltage, the screwdriver doesn't light on at all.
The house does have an RCD switch which works.
Assuming anything can be expected from the house wiring what could be the cause of the voltage on just some grounds?
There was a good comment if I want an engineering answer or how to fix this practically. I would like an engineering answer in the form of a summary of plausible theories in the EE jargon that would fit this behavior.