Buggered if I know why Garmin et al don't just include a USB port on the front of their devices.
Add a software modification to upload via the USB port using a simple delimited text file so just about anything can write the file, and a routine to compare the data with the internal database and allow corrections using the hardware's controls.
A "USB port", assuming you mean a USB slave and not the more complex USB controller function, would have to appear as a) one of the predefined USB devices e.g. storage, keyboard, mouse, in which case no driver is required on the laptop, or b) a non predefined device, in which case a laptop driver is required. Most USB ports on things like cameras or fancy phones can appear as storage devices so you just plug them in and the PC sees a new virtual hard drive and you just copy files to that.
I think this is what the current Jepp/Garmin x30W "flight plan migrator" solution is - except they use a flash cartridge as the transfer medium. This eases the verification requirements - not that GA avionics software undergoes all that much verification anyway
Anyone who tries this would have to ask themselves where they reckon any syntax checking takes place... Since the interface in question is not meant to be externally available, it could well be systemized under the assumption that any device transmitting on the interface will know what it is doing; i.e., hardly any syntax error protection. In that case, second worst case scenario is that a factory reset will bring the unit back to life again
There has to be some validation already, otherwise a bad crossfill connection would corrupt the destination GPS.
Obviously the PC software would have to be written carefully, but this is no different to what has been available on non-IFR GPS units for many years, with programs like Navbox or Flitestar. All these manage it OK right now. And anyway one looks at the displayed route on the GPS screen before one flies with it, so if your flight to Norwich shows a waypoint in Mongolia, it will be readily apparent
The crucial point is that one makes a lot more mistakes doing manual waypoint entry, which is why most pilots would go for a direct transfer if they could get it.
Incidentally, Chelton have been offering a flight plan loading method using a flash cartridge, for some years, so they must have come to the same conclusion as Garmin have much more recently... writing the flight plan onto the flash cartridge is a fairly robust way to do it. It will however b*gger up the CF connector pretty fast; in normal use you swap it just every 28 days but if you fly a lot you will end up swapping it before every flight. CF connectors are too flimsy for this purpose - but so are all USB connectors.
If the connector was a serial port then it could be remotely mounted, and easily replaced if damaged.
Which brings us to bluetooth or similar, with all the software reliability / compatibility issues. I have a very nice Thinkpad laptop on which Bluetooth stopped working after SP3 trashed it - even though I restored the HD from an image. No bluetooth...