Hummmmm.....
I'm not sure that I completely understand the issue. This completely depends on the method which the glass supplier are using to colour the glass. If they are using nanoparticles mixed into the glass matrix (i.e adding TiO2 into the batch pre-melt) then there would equally be as much light bleed into the rear side as there would be around the edges. If they are using a coating, (which from all accounts they are not) it would depend if they are using a CVD method or some form of in-line coating. In the case of an in-line they would have some low level light bleed due to the light refracting off the coated surface and coming out of the ends. If it is CVD chances are that they is coating on 5 of the 6 sides - hence no light bleed.
The problem is not easy to solve - since if they are using inclusions (which is my bet), they would need to increase the amount of TiO2 in the glass matrix which would then alter the other properties (strength, crack resilience, etc) making it more brittle.
The problem does not exist with black as it is an absorber, thus the amount of light propegating through the glass would need to be substantially higher in order to cause a problem.
Anyway - my thoughts.