These hand signals only make sense in cases where the attacker already has physical (though not mental) control over the victim and the crime is not immediately obvious to witnesses.
I think this case would actually be one of the easiest for the attacker to prevent with prior knowledge of the hand signal. Most people hardly ever look at the occupants of passing cars and only subset of them would have the knowledge and presence of mind to interpret the signal and not dismiss it anyway. So, a victim would have to have their fist pressed up against the glass for an extended period of time. The attacker, sitting right next to the victim, would catch that in their peripheral vision pretty quickly.
I also take offence to OP's suggestion that reasoned arguments critiquing the practice be 'automatically' moved to the bottom of the thread. As though such arguments can never be made in good faith.
Yes, there are clearly some situations where it has worked in the past, which is a good thing. But, we shouldn't start a massive public awareness campaign and create courses on it in schools because its aggregate effect would likely be small relative to other possible interventions.
Now, you may argue that you never suggested such a thing and that'd be a fair retort. However, then what is the potential aggregate impact such a programme? I'd argue very small for the reasons I've outlined previously. So, it be better to raise awareness and lobby for interventions to kidnapping and human trafficking with proven efficacy.