People believe many things that aren’t proven, even in science. For example, much of computer science is based on the belief that P is not equal to NP, even though that hasn’t been proven. All of math and science includes certain assumptions that must be made in order to make progress. So saying that it is being overly-critical to require “proof in a lab” to believe in God (or ghosts) is a red herring. People don’t require that level of proof to believe things.
However, testimony about ghosts is un-compelling – not because it isn’t “proven in a lab”, but because there is no clear tangible evidence to corroborate that testimony. Also, testimony about ghosts is wildly inconsistent.
In my opinion, testimony about god is also uncompelling for the same reasons. It makes zero sense to me that a god who wishes to be revered would leave no clear and tangible evidence of his existence. It also makes no sense that there would be over 4000 religions in the world, and god expects us to believe the testimony from one of them while discounting the testimony from the other 3999 – especially when most people in the world are guaranteed to not have even been exposed to that one or its adherents’ testimony.
Further, using the concept of “faith” to circumvent the above reasoning is something that I find disturbing. Being asked to believe something so important, specifically in the absence of evidence or logical support, is far more easily explained by it being a historical mechanism for people to control other people. And for that, there is plenty of corroborating evidence.