Actually there have been many tests of GI done with people who have diabetes, and many of the GI rankings in the University of Sydney data base are from tests done on people with diabetes. ( you can actually check each item to see who tested it, the number and type of people tested and when it was tested)
It's a ranking and in a comparison of 20 foods* tested with non diabetics and diabetics the correlation between them was 0.94 (over 0.9 is considered high)
if most non diabetic people get much higher glucose spikes with fluffy white rice than with lentils then so will a person with T2 diabetes (though the person with diabetes may have higher overall levels than the person without)
Portion control is important and that's where the GL comes in because it takes into account both the GI and the amount of carbs involved. It's no good thinking that pasta (cooked al dente) has a low GI and so eat the enormous bowl full that is often served in the UK (not so much in Italy!)
It's really a tool to distinguish between varieties of carb containing food. The emphasis recently has moved towards dietary patterns that tend to have an overall lower GI; so diets like the traditional Greek Mediterranean diet; In general minimally processed foods are lower in GI .
There is a world of difference for example in the GI of some whole barley or wheat cooked in a casserole or even as a porridge than the same grain finely milled flour and stuck into a 'whole grain' bread . (in the bread stakes coarser, denser and sourdough helps lower GI; ie more like peasant breads than typical supermarket offerings)
http://www.glycemicindex.com/ (the newsletters often have helpful information )
*
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2584181 This paper also has 2 supplements, lists of foods that have been tested. the first is in 'normal' the second includes those tested on PWD.