Returning to the subject of grounds in VFD systems, it is basic good practice to install a continuous copper ground conductor between the motor frame and the drive frame. This is not a safety issue but rather a way of preventing false drive faults (especially ground faults) due to different ground reference voltages. This ground conductor will be referenced to earth ground at the drive end but does not need to be referenced to earth ground at the motor end. In practice, the motor almost always is grounded too which produces a potential ground loop. As skogsgurra mentions, ground loops are overrated in power systems and, at least in the case of VFD's and motor leads, can be ignored.
If you have a ground shield or screen around your motor leads, in my view, the best practice is to follow signal and communication screening practice and ground the shield at the source end only--that would be the VFD end. The issue with the screen is entirely about suppressing EMI/RFI noise and ground loop currents are better avoided.
Be warned, however, that screening motor leads can lead to grievous common mode noise problems if your power supply system to the drive is floating. The only reliable solution I have found in those cases is the installation of a drive isolation transformer with a grounded center wye secondary. Not a pleasant solution if your customer was not expecting the extra cost and space requirements.
I hear that some reactor manufacturers are developing simpler, cheaper common mode noise solutions but I haven't used them and can't comment on their effectiveness. If anyone has experience with these newer devices, please post on their effectiveness.