Indeed, as you mentioned, some early swing-wing concepts had a complex mechanism that translated the root forwards, the idea being to keep the 'mean aerodynamic chord' (i.e. the chord at about 40% semi-span) roughly fixed relative to the aircraft. The concern was as much control as it was stability, but it was indeed related to movement of the aerodynamic centre relative to the cg.
In practice, the complexity hasn't been needed on any swing-wing I'm aware of, mainly because:
1. There is some corresponding aftwards cg motion when the wings sweep aft, which helps a bit.
2. There are changes to the downwash flowfield around the tail as a result of the wing sweeping closer to the tailplane. IIRC the Tornado tail was positioned quite carefully to take advantage of this.
3. The relative tail contribution to stability is actually reduced - if you think of the tailplane volume design coefficient:
VT = tail area * tail arm to wing AC / (wing chord * wing area)
you'll note that the tail arm decreases as the wing sweeps back, and the wing chord (streamwise) increases.
Since the overall aircraft stability is a combination of the wing-body and tailplane contributions, some reduction in the latter offsets the increase due to the aft motion of the wings.
4. the wing also becomes less efficient when swept in terms of generating lift, so while the AC is moved back, it's nothing like as bad as if the forward-swept wing were translated aft
and finally (though there are no doubt others)
5. the tail may be sized for low speed trim requirements (e.g. Tornado's very complex and powerful flap system no doubt requires a lot of trim authority at the tail) - in which case you may have tail power to spare at high speed.
Obviously it makes the design more complex - more configurations to worry about, especially if you have some kind of autosweep and dont have fixed positions. But it's not really much different to the design process than having a lot of flap positions - just another bunch of design points, another bunch of lines on the sizing diagrams, etc. It just so happens that designers, using some or all of the above factors, have managed to get a solution that didn't require forward translation of the root.
(As an aside, IIRC the F-14 wing gloves were supposed to be for AC position control, but were found to be unnecessary and were locked in for later operations)