They're magnetic, so if you have two conductors with equal current flowing in opposite directions the magnetic fields cancel out and you will read zero.
But there is no problem measuring current on a cable that is covered in plastic, or a cable that contains other conductors that aren't connected to anything.
You can even measure the total current of two separate circuits by feeding through the sensor either, the positive from both circuits, or the negative from both circuits. If you happened to feed the negative from one circuit and the positive from the other circuit through you would read the difference in the current of the two circuits.
Make sense?
One side note: Hall sensors that read AC current don't do any processing with the data, on the output you will get an AC waveform that represents the current, not a steady voltage. You need to do your own RMS calculations if you want to measure AC current.
For DC current through the sensor though, you get a DC voltage that you can read directly to get the DC current on the conductor.
Quite often the output is like 2.5V at zero Amps and moves up to 5V or down the 0V depending on the direction of the current.