Machines are rotating devices, according design the permissible air gap is minimum as possible yet if we use oil, though reluctance may reduce yet the friction offer would be more and efficiency would not increased much.

I would think that the permeability of mineral oil is about the same as that of air (or free-space), combined with, as Bhupendra pointed out, increased friction (more viscous fluid), it would be a net loss.

If instead of plain oil, a ferrofluid were used, and the relative surface speed were low (small device or low RPM) there may be some gain. Not sure how stable or abrasive a ferrofluid might be over the device lifetime.

Usually induction motors used in oil drilling rig are filled with oil in the air-gap.

Submersible water pump motor too oil filled yet the efficiency is not increasing much. This are the special purpose motor and are costly, repair cost is also high, not suitable for industrial use.

I hadn't thought about this issue and it is interesting. I have two questions:

1. heat transport from the rotor? Would the heat of even a slow moving rotor eventually create a train of involved gas products that would potentially reduce friction? This is an effect much like boundary layer control.

2. vibration transmission to other parts? It would seem that the air space provides a natural insulator against acoustic impacts on other system parts that a incompressible fluid might not.

I don't have much time to really look at these speculations.

With oil filled elect. machine cooling may be better, running

may better hence life of machine could increased but not the efficiency effect much.

Viscosity of air is 17 e−6 Pa·s, oil has 50-200 e+3, it means machanical loss due to friction will significally increased. In large turbogenerators hydrogen (8e-6) is used instead od air, to reduce rotor friction loss.