In an AC motor where you use a laminated magnetic core to shield the induction process from eddy currents, the field lines that run through it result in core loss through heat. Is all of the energy lost through this mechanism, can any of it be recovered like in regenerative braking in cars? And if all of the field lines don't dissipate through the resistance in the material, where do they go?
A very small amount of the energy goes into vibrating the core. This will cause a heavily-loaded transformer to audibly hum.
You may find this article http://www.electricaleasy.com/2014/04/transformer-losses-and-efficiency.html useful. Note that, for a well-designed, well-constructed transformer, efficiencies are already very high, and heat losses are not directly recoverable, so trying to scavenge any stray field energy doesn't make much sense.
Energy is not "lost" - it just changes how it is expressed / stored, and may become less useful (harder to convert to another, more useful form). That's entropy for you.
For example, any time you have a source of heat, that source can be used to drive a heat engine and recover some energy. Since motors are usually designed to produce little heating, the temperature difference for the heat engine is likely to be small - making it less efficient.
But sometimes the heat can be used directly for some useful task. For example, the motor of an aquarium pump does not only circulate the water through the filter, but it also acts to heat the water (it is immersed) so while you keep the motor cool, you keep the water warm. Win-win.
Losses in ferromagnetic materials
There are two mechanisms which produce losses in ferromagnetic materials. One of them is hysteresis, the other is eddy currents.
Hysteresis losses occur while the magnetic dipoles rotate, meaning while you change the direction of magnetization. You can look at it through single dipoles or through Weiss domains. When you change the field direction and intensity, they will reorient which makes them move on the hysteresis loop. Part of the energy is lost, the other part of the energy is stored as reoriented domains.
They are induced currents in the laminations which will create losses, which depend strongly on the conductivity of the material in the direction in which they want to flow.
All losses(that's why we call them that way) are converted to heat(or vibration), they are transformed to another form of energy. You can't turn heat into a magnetic field easily, for this process you will either need a thermoelectric generator which is inefficient and has very very low power(while it costs a fortune), or you use a thermodynamic cycle to convert it to mechanical work, which in turn can be used to drive a generator. The heat you generate is low quality energy/heat(from an energy conversion point of view). It is like trying to get a turbine going with a big pool of lukewarm water, it makes no practical sense. That's why we create highly energetic (high pressure) steam in electric power stations.
All in all, losses are lost. If they weren't lost we would call them differently. If the energy is stored in the system temporarily we call it stored energy, in AC systems this can be interpreted as reactive power(look it up). Generally you try to minimize losses, as it brings a lot more than energy recovery for a lower price.