Exactly, and I agree completely. In this case, since few, if any, non-diabetics were subjects, except in a few cases where the control group were composed of some non-diabetics, the data does not translate well into the ramifications for the non-diabetic. What the literature does seem to indicate is that there is no negative feedback loop with regards to insulin and ALA use. From what I can find, there is a dose limit of 600 mg and normalization of glucose. That is, at a dose higher than 600 mg there is neither increased or decreased benefit (short of doses approaching the LD50). So when discussing the effects of high dose ALA, the insulin question is pointless comparing diabetics to non-diabetics.
There is one study of 1800 mg per day that did not report any unusual side effects. For this study it confirmed that with increased dose of ALA there is another perhaps unknown mechanism via which ALA is functioning. In this case, assuming there are no other pathological conditions present in the study group, it does bring up the question of decreased glycation and increase myocyte uptake of glucose. Some ridiculously incomparable animal studies have shown this but it will be nice to see it with higher primates, or even some Italian human studies.