Edit: I realize now that measuring the difference in resistance is not practically easy as the resistivity difference isn't significant, and there are numerous other measurement challenges including surface areas for contacts creating a significant enough difference for this method to work. Other methods may be more suitable.
If it's a cuboid, then I think a simpler way would be to measure its resistance.
- Use Suitable contacts
- Measure resistance
- Use R=ρL/A and find out resistivity ρ.
- Either use the table found here and check if it matches up
You can also use similar method for other properties like
-Thermal conductivity: Heat one end and monitor how quickly the other end's temperature rises, etc and compare to expected value or just another gold bar of similar dimension(which shouldn't be too difficult)
-Ultrasonic pulse (like'Pulsar' mentioned)
-A similar technique can be used with other rf wave reflection analysis. The pattern reflected would differ from a solid gold bar.
-Another method that comes to mind is Electrodynamic suspension. You can suspend it in air, but the eddy current will be different because it is hollow, and thus the induced magnetic field strength will be different. For a given time-varying magnetic field function, the response would differ.
But I personally think that the resistance measurement is the easiest.
All these are non destructive, btw. (Not so much for the scammer.)