You can't get away from the fact that to get a given amount of power you have to burn a given amount of fuel. Nor can you get away from the fact that to burn a given amount of fuel you create a given amount of CO
2. It's just chemistry. Engine efficiency varies but not in orders of magnitude.
As Martin eludes to, the test isn't exactly representative of "normal" driving so you can tweak a car to appear very efficient whereas the reality is far from that. When the GM V6 was developed there was no need to.
Two things that stand out about the AJ-V8 are that it has an ali block, so probably warms up much quicker than the lump of pig iron that is the GM V6 block, and that it has VVT, so can probably be tuned to run more efficiently under light load, so it does better during the urban bits of the test cycle and the warmup cycle.
In reality, people drive for further than 10km and I don't doubt that if I had a 4.2 supercharged V8 I would burn more fuel

, and hence create more CO
2 than I do currently.
These figures and the tests that determine them are dreamed up by bureaucrats in Brussels who have no idea what happens under the bonnet of a car. No idea about anything at all, in fact. The car industry then has to manipulate their products to get round the obstacles placed in their path.
Kevin