I don't know about the Weims/lab situation, but on the GSDs, because it's not a hugely rare variation, I wouldn't think a breeder that has a bunch of them has been playing fast and loose with the breed.
GSDs have a huge amount of variation in the breed. It's a real chore for those who breed with some purpose, to hold together all the different directions that GSD genetics are going in. So, if any breeder loses sight of that for something as superficial as color, it wouldn't take long for the dogs to look poorly bred, and eventually not really even look all GSD. Honestly I'd imagine the same could be said of labs - eyes would be sharp to spot supposed similarities to Weims but for most I don't see how you could tell. Weims look so much like labs anyway.
Blue doesn't have to do with liver/chocolate, which is a different gene than black (ie, a dog can be black, or liver, but not both at the same time). The dilute gene acts on the locus which carries liver and black (B/b), so you can have a dilute black or dilute liver. It doesn't affect tan markings, although it does affect dominant yellow color (like Irish setters and golden retrievers and yellow labs). On yellow, however, it can be hard to spot.
The gene is recessive. That means if you get two parents that are expressing it (they have two copies of it), 100% of their pups will also be dilutes. So it's very easy to increase rapidly in a population once you get a few separate individuals with the gene. The problem is, a recessive trait is also one that is randomly expressed AND may be linked in some way to one or more OTHER recessive traits. These may be relatively benign like a more feminine head, or even a more collie-like head, more moderate ears, a lighter frame - or the opposite. Or, they may be unpleasant traits like a structural weakness or a disease.
Another good example of this can be seen in the breeding of a fellow named Gary Ericsson. He was a breeder of working Border Collies and Aussies. His standards were intensely high. His Aussies were very focused and could do long outruns and distance field work as well as a good Border Collie, in addition to being good at the things Aussies are good at - going in hard when needed and not taking "no" for answer. If you saw one of his dogs working it would be hard to tell it from a Border Collie. They work with a low head, crouching, using their eyes to hold stock and cut off their escape instead of bouncing around from side to side. I've seen a couple of Aussies with strong Hanging Tree lines and they are amazing.
People started yelling that he had combined Border Collies and Aussies. They pointed to physical characteristics in his Aussies that were "like BCs." His dogs tended to be solid colors instead of merles. They were lighter boned, longer bodied, and some had upright ears.
This was silly. If you select for Border Collie like characteristics in dogs that are already very similar to BCs, then of course they'll start resembling BCs in many ways. Especially a more or less recessive trait like eye. The same thing has been done in rough collies (NZ Heading Dogs) and working Beardies with the same result - dogs that look and work for all the world like Border Collies.
Gary Ericcson did eventually create his own breed, but it wasn't a BC/Aussie cross - it was BC/cur dog. That breed, the Hangin Tree Cowdog (HTC) is just shy of achieving true breed status - a stable gene pool is in place, and there were three separate registries that just voted to combine. I have a couple friends with HTCs and it should be fun to watch this happen. I may go visit this guy's ranch sometime next year.
The other factor in basing your line on a recessive trait, is the founder effect - such dogs will set characteristics in a line very quickly and you won't be able to get rid of it, short of DNA testing. Quarter horses are living that nightmare still, as are the rough collie people.
Merle is a different matter altogether. If a breed HAS merle in the genetics, then a merle parent must be used to create a merle line. It's dominant so there always has to be one merle parent. There can't be two, as with black (also dominant), to increase the odds of a merle, because it's a lethal dominant and pups with two genes are usually severely handicapped or don't make it post partum. Some do it anyway as the double merle pups are used to increase merle in the line (ick), but it's not Nice. At least I don't think so.
I just fostered and placed a darling little double merle sable Border Collie and her new owners took her on knowing what they were getting into - they wanted a "project dog". Still, it's heartbreaking to hear of their constant trips to the vet - thankfully her life expectancy is excellent, as is her quality of life thanks to their wonderful care of her, but she's deaf, has limited eyesight, has a deformed spine and hip, and has skin and immune system troubles.
Basically, it's my personal belief that breeding strictly for color is not responsible. Dogs aren't fashion statements.